Ad ~ "' c~ . T" ts
This book has been a long time in the making . Its creation has been a Widely distributed cognitive process. I wish to thank first those who provided me the opportunity to make the observations on which this work is based. I am grateful to the crews of the Palau (a pseudonym ) and all the other ships I sailed upon . The commanding officer and the navigator of the Palau merit special recognition for allowing me to work aboard their ship . I am especially grateful to ' the quartermaster chief and the men of the Palau s Navigation Department for working with me and sharing their working lives so generously . Although I will not name them here or in the text , they know who they are and I am grateful to them . James Tweedale , then Technical Director of the Navy Personnel Research and Development Center, generously supported the early phases of the research as an independent research project . Additional ' support was provided by the Office of Naval Research s Division of Psychology and Personnel Training under the guidance of Susan Chipman and Michael Shatto. My supervisor and colleague at NPRDC, James Holian , provided a great working environment for me and helped me to organize my thinking in the early stages. Barbara Morris and Michael Goelier helped with the transcriptions and coding of the data. Colleen Siefert worked with me - as a postdoc , made observations on another ship , and co authored portions of the discussion of learning from error . I thank the John D. and Catherine T. Mac Arthur Foundation for a five -year foundation fellowship that permit ted me to work on this material when no suitable institutional setting existed . Perhaps more important , the fellowship gave me the courage to follow ideas that lay outside the mainstream . Over the years in which this work developed , I profited from my involvement in the cognitive science community at the University of California at San Diego. I am especially grateful to Donald Norman , who shared many ideas with me as we ran a research laboratory and taught courses together . I am also grateful to AcknowledgementsI
' Aaron Cicourel , Roy D Andrade , Rik Belew , Mike Cole, and Yrjo Engestrom for helping me think through these ideas. The preparation of the book was facilitated by the helpful comments of Bambi Schieftelin , Jacques Theureau , Everett Palmer , Nick Flor , and Christine Halverson . My greatest debt is to my wife , Dona, who provided encouragement , support , great meals, and editorial assistance throughout the project . IbtDdl ~" "
The seed from which this book grew was planted in November 1980, when I spent most of a day on the navigation bridge of a u .s . Navy ship as it worked its way in from the open North Pacific , through the Straits of Juan de Fuca, and down Puget Sound to Seattle. I was aboard the ship to study what the operators of its steam propulsion plant knew and how they went about knowing it . I had spent most of the preceding week down in the bowels of the ship , observing engineering operations and talking to the boiler ' technicians and machinist s mates who inhabited that hot , wet , noisy tangle of boilers , pumps , and pipes called the engineering ' spaces. I ll admit to having felt a little claustrophobic after all that time spent below the water line , where there is no night or day and the only evidence of being at sea is the rhythmic tipping of the deck ' plates and sloshing of water in the bilge below one s feet as the ship rolls in the swell . A chief boiler technician confided to me that in 21 years on Navy ships he had never yet been on deck to experience ' either of those two most romantic seafaring events, a ship s arrival at or departure from a port . I resolved , therefore , to take my last few hours aboard this ship on the navigation bridge , where I could see out the windows or even go out on the bridge wing to get a breath of cold fresh air . My professional rationalization for being on the bridge was that there I would be able to observe the process that generates the fluxry of engine commands that always taxes the engineering crew when the ship nears the dock . And I did make a detailed record of all engine knd helm commands given in the 75 minutes from the time the - engines were first slowed until they were secured there were 61 in all . But what really captured my attention was the work of the navigation team. Three and a half years later , the project that became this book began in earnest. In the SI)mmer of 1984, I was still working for the Navy Personnel Research and Development Center in San Diego as a civilian scientist with the title Personnel Research Psychologist . - By then I had participated in two successful and well known Inb' oduction d
projects . With these success es came the freedom to conduct an independent research project . I was given carte blanche to study whatever I thought was of most interest . I chose to study what I was then calling naturally situated cognition . Having a research position in a Navy laboratory made it possible for me to gain access to naval vessels, and my longtime love of navigation and experience as a racing yacht navigator made it easy for me to choose navigation as an activity to study afloat. I talked my way aboard a ship and set up shop on the navigation bridge . At the time , I really had no notion what an ideal subject navigation would turn out to be. When I began, I was thinking in terms of the naturally situated cognition of individuals . It was only after I completed my first study period at sea that I realized the importance of the fact that cognition was socially distributed . A little earlier , I had been asked to write a book describing what is in cognitive anthropology for the rest of cognitive science. I began that project , but after I became disillusioned with my field I lost interest in it . The choice of naturally situated cognition as a topic came from my sense that it is what cognitive anthropology really should have been about but largely had not been. Clifford Geertz " " (1983) called for an outdoor psychology , but cognitive anthro - pology was unable or unwilling to be that . The respondents may have been exotic , but the methods of investigation were largely borrowed from the indoor techniques of psychology and linguistics . When cognitive and symbolic anthropology split off from social anthropology , in the mid 1950s, they left society and practice behind . As part of the cognitive revolution , cognitive anthropology made two crucial steps. First , it turned away from society by looking inward to the knowledge an individual had to have to function as a " member of the culture . The question became What does a person " have to know ? The locus of knowledge was assumed to be inside the individual . The methods of research then available encouraged the analysis of language. But knowledge expressed or expressible in language tends to be declarative knowledge . It is what people can say about what they know . Skill went out the window of the " " white room . The second turn was away from practice . In the quest to learn what people know , anthropologists lost track both of how people go about knowing what they know and of the contribution of the environments in which the knowing is accomplished . Perhaps these narrowing assumptions were necessary to Introduction xl
get the project of cognitive anthropology off the ground . I will argue that , now that we are underway as a discipline , we should revoke these assumptions . They have become a burden , and they prevent us from seeing the nature of human cognition . In particular , the ideational definition of culture prevents us seeing that systems of socially distributed cognition may have interesting cognitive properties of their own . In the history of an- thropology , there is scarcely a more important concept than the division of labor . In terms of the energy budget of a human group and the efficiency with which a group exploits its physical environment , social organizational factors often produce group properties that differ consider ably from the properties of individuals . Clearly , the same sorts of phenomena occur in the cognitive domain . Depending on their organization , groups must have cognitive properties that are not predictable from a knowledge of the properties of the individuals in the group . The emphasis on finding and " " " " describing knowledge structures that are somewhere inside the individual encourages us to overlook the fact that human cognition is always situated in a complex sociocultural world and cannot be unaffected by it . Similar developments in the other behavioral sciences during the cognitive revolution of the late 1950s and the 1960s left a troubled legacy in cognitive science. It is notoriously difficult to generalize laboratory findings to real-world situations . The relationship between cognition seen as a solitary mental activity and cognition seen as an activity undertaken in social settings using various kinds of tools is not at all clear . This book is about softening some boundaries that have been made rigid by previous approach es. It is about locating cognitive activity in context , where context is not a fixed set of surrounding conditions but a wider dynamical process of which the cognition of an individual is only a part . The boundaries to be softened or dissolved have been erected, primarily for analytic convenience , in social space, in physical space, and in time . Just as the construction of these boundaries was driven by a particular theoretical perspective , their dissolution or softening is driven by a different perspective - one that arose of necessity when cognition was confronted in the wild . " " The phrase cognition in the wild refers to human cognition in its natural habitat - that is, to naturally occurring culturally constituted " " human activity . I do not intend cognition in the wild to Introduction xlv
' " " be read as similar to Levi -Strauss s pensee sauvage, nor do I intend ' it to contrast with Jack Goody s (1977) notion of domesticated mind . Instead , I have in mind the distinction between the laboratory , where cognition is studied in captivity , and the everyday world , where human cognition adapts to its natural surroundings . I hope to evoke with this metaphor a sense of an ecology of thinking in which human cognition interacts with an environment rich in organizing resources. The attempt is cultural in nature , giving recognition to the fact that human cognition differs from the cognition of all other animals primarily because it is intrinsically a cultural phenomenon . My aim is to provide better answers to questions like these: What do people use their cognitive abilities for ? What kinds of tasks do they confront in the everyday world ? Where shall we look for explanations of human cognitive accomplishment ? There is a common misconception among cognitive scientists , especially those who do their work in laboratory settings , that research " " conducted outside the laboratory is necessarily applied work . I will argue in what follows that there are many excellent reasons to look at the " real world " that are not concerned with hoped -for applications of the research findings (although funding sponsors often like to think in those terms). Pure research on the nature of real cognitive practices is needed. In this book , I emphasize practice not in order to support a utilitarian or functionalist perspective but because it is in real practice that culture is produced and reproduced . In practice we see the connection between history and the future and between cultural structure and social structure . One of my goals in writing this book is to make clear that the findings of pure research on cognition in the wild should change our ideas about the nature of human cognition in general. This is not news to anthropologists , who have been doing pure research in the form of ethnography for decades. This book is an attempt to put cognition back into the social and cultural world . In doing this I hope to show that human cognition is not just influenced by culture and society , but that it is in a very fundamental sense a cultural and social process. To do this I will move the boundaries of the cognitive unit of analysis out beyond the skin of the individual person and treat the navigation team as a cognitive and computational system. " " Chapter 1, Welcome Aboard , attempts to locate the activity of ship navigation in the larger world of modem life . It weaves to- Inb' oduction IV
gether three journeys : a movement through physical space from the " " street to the ship , a movement through social space from civilian to military life , and a movement through conceptual space from everyday notions of wayfinding to the technical domain of navigation . Both the researcher and the reader must make these journeys to arrive at the activity of navigation as practiced on the bridge of a Navy ship . Military ranks and the ways in which military identities are formed are presented here because these things affect in - ' dividual s relationships to their work . An important aspect of the larger unit is that it contains computational elements (persons) who cannot be described entirely in computational terms . Who they talk to and how they talk to one another depend on these social organi - zational factors. This chapter also contains a discussion of the relationship of the researcher to the activity under study . (The name of the ship and the names of all the individuals mentioned in the book are pseudonyms . All the discourses reported , whether standing alone in transcript form or embedded in narrative passages were transcribed directly from audio recordings of actual events.) Having taken navigation as it is performed by a team on the bridge of a ship as the unit of cognitive analysis , I attempt in chapter " " 2, Navigation as Computation , to apply the principal metaphor - of cognitive science cognition as computation - to the operation of this system. I should note here that in doing so, I do not make any special commitment to the nature of the computations that are going on inside individuals except to say that whatever happens there is part of a larger computational system. This ' chapter describes the application of David Marr s notions of levels of analysis of cognitive systems to the navigation task and shows that , at the computational level , it is possible to give a single description of the computational constraints of all known technical forms of human navigation . A comparison of modem Western navigation with navigation as practiced in Micronesia shows that considerable differences between these traditions lie at the representational /algorithmic level and at the implementationallevel . A brief historical review of the development of modem navigation shows that the representational and implementational details of contemporary practice are contingent on complex historical proc - esses and that the accumulation of structure in the tools of the trade is itself a cognitive process. - Chapters 3 5 explore the computational and cognitive properties of systems that are larger than an individual . The issues addressed Introduction xvi
in these chapters concern how these larger systems operate and how their cognitive properties are produced by interactions among their parts . " " Chapter 3, The Implementation of Contemporary Pilotage , describes the physical structures in which the navigation computations are implemented . This chapter elaborates a conception of computation as the propagation of representational state across a variety of media . This view of computation permits the use of a single language of description to cover cognitive and computa - tional process es that lie inside and outside the heads of the practitioners of navigation . The first section of this chapter describes the " " fix cycle as a cognitive process. The second section describes how navigation tools are used and how local functional systems composed of a person in interaction with a tool have cognitive properties that are radically different from the cognitive properties of the person alone. The third section discuss es the ways in which the computational activity can be distributed through time by precomputing not only partial results but also the means of computation . I show here how the environments of human thinking are not " " natural environments . They are artificial through and through . Humans create their cognitive powers by creating the environments in which they exercise those powers . This chapter concludes with a discussion of the relationship between the cognitive properties of the individuals performing a task and the cognitive properties of the system in which they participate . " " Chapter 4, The Organization of Team Performances, moves the boundaries of the unit of analysis even further out to consider the cognitive properties of the team as a whole . Here I note some of the problems that are encountered when cognitive activities are distributed across the members of a group . It is not the case that two or more heads are always better than one. This chapter describes the structures and process es involved in the group performance of the navigation task. The first section follows through on the application ' of Marr s concepts of computation to the navigation activity and discuss es the properties of the activity as an explicitly computa - t.ional system. The second section presents a problem in work organization encountered by the navigation team and shows why it is often difficult to the that individual action . apply concepts organize to the organization of group action . The final section shows how the members of the navigation team form a flexible connective tissue that maintains the propagation of representational state in the face of a range of potentially disruptive events. Inb' oduction xvi
" " Chapter 5, Communication , continues the theme of chapter 4 but looks at communication in more detail . It asks: How is it that patterns of communication could produce particular cognitive properties in a group ? The chapter begins with a discussion of features of communication observed in the navigation team and their ' effects on the Team s computational properties . These observations lead to some simple hypotheses about the ways in which patterns of communication might affect the computational properties of a group . These hypotheses are explored using a computer simulation of communities of connectionist networks . The simulations lead to the surprising conclusion that more communication is not always better . - Chapters 6 8 concern learning or change in the organization of cognitive systems at several scales. " " Chapter 6, The Context of Learning , is a bridge between the descriptions of ongoing operations provided by the previous chapters and the descriptions of changes in the nature of ongoing operations provided by the following chapters . It describes the context in which novice navigators become experts . This chapter is an attempt to examine both the work that the system does in order to scaffold learning by practitioners and the opportunities for the development of new knowledge in the context of practice . Whereas in chapter 6 I deal with the observable contexts surrounding " " learning , in chapter 7, Learning in Context , I Uy to dissolve the boundaries of the skin and present navigation work as a system of interactions among media both inside and outside the individual . I look at learning or conceptual change as a kind of adaptation in a larger dynamical system. This chapter presents a functional notation and a framework for thinking about learning as local adaptation in a dynamic system of coordinations of representational media . " " Chapter 8, Organizational Learning , returns the focus to the larger unit of analysis : the team as a whole . It presents a case study of an incident in which the navigation team was forced to adapt to changes in its information environment . The analysis presented here examines a particular incident in which the microstructure of the development of the navigation practice can be seen clearly . It is an attempt to show the details of the kinds of process es that must be the engines of cultural change. " " Chapter 9, Cultural Cognition , attempts to pull the preceding chapters together into a coherent argument about the relationships of culture and cognition as they occur in the wild . I attempt first Inuoduction xvi
to illustrate the costs of ignoring the cultural nature of cognition . I argue that a new framework is needed to understand what is most characteristically human about human cognition . I Ii order to construct a new framework , the old one must be deconstructed . I therefore provide two readings of the history of cognitive science: a history as seen by the proponents of the currently dominant paradigm and a rereading of the history of cognitive science from a sociocultural perspective . The differences between these two readings highlight a number of problems in contemporary cognitive science and give new meanings to some of the familiar events in its history . 1 W~ W118Aboatd
N8 T8tive: A ~ After several days at sea, the U.SiS. Palau was returning to port , making approximately 10 knots in the narrow channel between Ballast Point and North Island at the entrance to San Diego Harbor . In the pilothouse or navigation bridge , two decks above the flight deck, a junior officer had the conn (ie ., was directing the steering of the ship ), under the supervision of the navigator . The captain sat quietly in his chair on the port side of the pilothouse watching the work of the bridge team. Morale in the pilothouse had sagged during two frustrating hours of engineering drills conducted just outside ' the mouth of the harbor but was on the rise now that the ship was headed toward the pier . Some of the crew talked about where they should go for dinner ashore and joked about going all the way to the pier at 15 knots so they could get off the ship before nightfall . " The bearing recorder had just given the command Stand by to " mark time 3 8 and the fathometer operator was reporting the depth of the water under the ship when the intercom erupted with the " voice of the engineer of the watch : Bridge , Main Control . I am ' losing steam drum pressure. No apparent cause. I m shutting my " throttles . Moving quickly to the intercom , the conning officer acknowledged " " : Shutting throttles , aye. The navigator moved to the ' " captain s chair , repeating : Captain , the engineer is losing steam on " the boiler for no apparent cause. Possibly because he realized that the loss of steam might affect the steering of the ship , the conning officer ordered the rudder amidships . As the helmsman spun the wheel to bring the rudder angle indicator to the centerline , he answered " " the conning officer : Rudder amidships , aye sir . The captain " " began to speak, saying Notify , but the engineer was back on the intercom , alarm. in his voice this time , speaking rapidly , almost " ' shouting : Bridge , Main Control , I m going to secure number two " boiler at this time . Recommend you drop the anchor ! The captain had been stopped in mid -sentence by the blaring intercom , but before the engineer could finish speaking the captain said, in a loud " " but cool voice , Notify the bosun . It is standard procedure on Chapter1 2
large ships to have an anchor prepared to drop in case the ship loses its ability to maneuver while in restricted waters . With the propulsion plant out , the bosun , who was standing by with a crew forward ready to drop the anchor , was notified that he might be ' called into action . The falling intonation of the captain s command gave it a cast of resignation or perhaps boredom and made it sound entirely routine . In fact, the situation was anything but routine . The occasional cracking voice , a muttered curse, or a perspiration -soaked shirt on this cool spring afternoon told the real story : the Palau was not fully under control , and careers and possibly lives were in jeopardy . The immediate consequences of this event were potentially ' grave. Despite the crew s correct responses, the loss of main steam put the ship in danger. Without steam, it could not reverse its propeller - the only way to slow a large ship efficiently . The friction of ' the water on the ship s hull will eventually reduce its speed, but the Palau would coast for several miles before coming to a stop. ' The engineering officer s recommendation that the anchor be dropped was not appropriate . Since the ship was still traveling at a high rate of speed, the only viable option was to attempt to keep the ship in the deep water of the channel and coast until it had lost enough speed to safely drop anchor . Within 40 seconds of the report of loss of steam pressure, the steam drum was exhausted . All steam-turbine -operated machinery came to a halt , including the turbine generators that produce the ' ship s electrical power . All electrical power was lost throughout the ship , and all electrical devices without emergency power backup ceased to operate. In the pilothouse a high -pitched alarm sounded for a few seconds, signaling an under -voltage condition for one piece of equipment . Then the pilothouse fell eerily silent as the electric motors in the radars and other devices spun down and stopped . Just outside the navigation bridge , the port wing pelorus operator watched the gyrocompass card in his pelorus swing wildly and then return to its original heading . He called in to the bearing " recorder standing at the chart table : John, this gyro just went " nuts . The bearing recorder acknowledged the comment and told " the pelorus operator that a breakdown was in progress: Yeah, I ' ' " know , I know , we re havin a casualty . Because the main steering gear is operated with electric motors , the ship now not only had no way to arrest its still -considerable W Alcnm A Aboard 3
forward motion ; it also had no way to quickly change the angle of its rudder . The helm does have a manual backup system, located in a compartment called aftersteering in the stem of the ship : a worm - gear mechanism powered by two men on bicycle cranks . However , even strong men working hard with this mechanism can change the angle of the massive rudder only very slowly . Shortly after the loss of power , the captain said to the navigator , " who was the most experienced conning officer on board , OK, ' " " Gator, I d like you to take the conn . The navigator answered Aye , " " sir and , turning away from the captain , announced : Attention in " the pilothouse . This is the navigator . I have the conn . As required , " the quartermaster of the watch acknowledged ( Quartermaster , " " " aye ) and the helmsman reported Sir , my rudder is amidships . The navigator had been looking out over the bow of the ship , trying " to detect any turning motion . He answered the helmsman : Very " well . Right 5 degrees rudder . Before the helmsman could reply , " the navigator increased the ordered angle: Increase your rudder " right 10 degrees. (The rudder angle indicator on the helm station has two parts ; one shows the rudder angle that is ordered and the other the actual angle of the rudder .) The helmsman spun the wheel , causing the indicator of the desired rudder angle to move to the right 10 degrees, but the indicator of the actual rudder angle " " seemed not to move at all . Sir , I have no helm sir ! he reported . Meanwhile , the men on the cranks in aftersteering were straining to move the rudder to the desired angle. Without direct helm control ' , the conning officer acknowledged the helmsman s report and sought to make contact with aftersteering by way of one of the " " phone talkers on the bridge : Very well . Aftersteering , Bridge . The " navigator then turned to the helmsman and said Let me know if " you get it back. Before he could finish his sentence, the helmsman " " responded , I have it back, sir . When the navigator acknowledged the report , the ship was on the right side of the channel but heading " far to the left of the desired course. Very well , increase your rudder " " to right 15. Aye sir . My rudder is right 15 degrees. No new " " " course given . The navigator acknowledged - Very well - and " then , looking out over the bow , whispered Come on, damn it , " swing ! Just then , the starboard wing pelorus operator spoke on the " ' phone circuit : John, it looks like we re gonna hit this buoy over " here. The bearing recorder had been concentrating on the chart ' " " and hadn t quite heard . Say again he requested. The starboard wing pelorus operator leaned over the railing of his platform to Chapter1 4
watch the buoy pass beneath him . It moved quickly down the side of the ship , staying just a few feet from the hull . When it appeared that the Palau would not hit the buoy , the starboard wing pelorus " ' " operator said Nothin ; that ended the conversation . The men inside never knew how close they had come. Several subsequent " " helm commands were answered with Sir , I have no helm . When asked by the captain how he was doing , the navigator , referring to " their common background as helicopter pilots , quipped First time " " " I ever dead-sticked a ship , captain . (To dead-stick an aircraft is to fly it after the engine has died .) Steering a ship requires fine ' judgements of the ship s angular velocity . Even if helm response was instantaneous , there would still be a considerable lag between ' the time a helm command was given and the time when the ship s response to the changed rudder angle was first detectable as the movement of the bow with respect to objects in the distance . Operating with this manual system, the navigator did not always know what the actual rudder angle was, and could not know how long to expect to wait to see if the ordered command was having the desired effect. Because of the slowed response time of the rudder , the navigator ordered more extreme rudder angles than usual , causing the Palau to weave erratically from one side of the channel to the other . Within 3 minutes , the diesel-powered emergency generators were brought on line and electrical power was restored to vital systems throughout the ship . Control of the rudder was partially restored , but remained intermittent for an additional 4 minutes . Although the ship still could not control its speed, it could at least now keep itself in the dredged portion of the narrow channel . On the basis of the slowing over the first 15 minutes after the casualty , it became possible to estimate when and where the Palau would be moving slowly enough to drop anchor . The navigator conned the ship toward the chosen spot. About 500 yards short of the intended anchorage, a sailboat took a course that would lead it to cross close in front of the Palau . Normally the Palau would have sounded five blasts with its enormous horn to indicate disagreement with the actions taken by the ' other vessel. However , the Palau s horn is a steam whistle , and without steam pressure it will not sound . The Navigation Department has among its equipment a small manual foghorn , basically a bicycle pump with a reed and a bell . The navigator remembered Welcome Aboard 5
this of and instructed the of the deck to leave piece gear keeper log his find the manual horn descend two levels to the post , , flight deck take the horn out to the bow and sound the five , , warning blasts . The of the deck ran from the keeper log pilothouse , carrying
- a walkie talkie to maintain communication with the . The bridge ' the for the deck s address captain grabbed microphone flight public " " and asked Can hear me on the deck ? Men below system you flight " on the deck turned and waved at the . Sailboat up pilothouse ' Palau s bow be advised that I am not . . . I have no . crossing power " You cross at own risk . I have no . this time the hull your power By , of the sailboat had under the bow of the and disappeared ship only its sails were visible from the . In the the pilothouse foreground , men on the deck were now to the bow to watch the flight running
collision . Meanwhile the of the deck had impending , keeper log run down two of stairs from the base of the island flights , emerged ,
and across the 100 that between begun sprinting nearly yards lay the island and the bow . Before he was to his it was halfway goal ,
clear that the time he would reach the bow the from the by signal
horn would be . The turned to a officer meaningless navigator junior " - who was a walkie talkie and exclaimed tell him to holding Just " the sucker down and hit it five times ! The was put message
and the five feeble blasts were sounded from the middle of passed ,
the deck . There is no to know whether the was flight way signal
heard the sailboat which then was ahead of the by , by directly
Palau and so close that the of its mast was visible from the only tip
. A few seconds later the sailboat still pilothouse , emerged , sailing ,
from under the starboard bow . The of the deck continued keeper log
to the bow to take a there in case other up position warnings
were . required
- five minutes after the and more Twenty engineering casualty
than 2 miles from where the wild ride had the Palau was begun ,
to anchor at the intended location in water outside brought ample just
the bounds of the channel . navigation
The safe arrival of the Palau at anchor was due in to large part
the of the crew the navigator exceptional seamanship bridge , especially
- . But no individual on the alone neither single bridge acting
the nor the nor the chief captain navigator quartermaster supervising
- the team could have control of the navigation kept ship
and it to anchor . kinds of were brought safely Many thinking
to this task . Some of them were in required perform happening Chapter1 8
parallel , some in coordination with others , some inside the heads of individuals , and some quite clearly both inside and outside the heads of the participants . This book is about the above event and about the kind of system in which it took place. It is about human cognition - especially human cognition in settings like this one, where the problems that individuals confront and the means of solving them are culturally structured and where no individual acting alone is entirely responsible for the .outcomes that are meaningful to the society at large. Gaining access to this field site required me, as an ethnographer , to make three journeys at once. In this first chapter I will try to weave them together , for the reader will also have to make these journeys mentally in order to understand the world of military ship navigation . The first is a journey through physical space from my home and my usual workplace to the navigation bridge of the Palau . This journey took me through many gates, as I moved from the street to the military base, to the ship , and within the ship to the navigation bridge . I will try to convey the spatial organization of the setting in which navigation is performed . The second journey is a trip through social space in which I moved from the civilian social ' world past the ship s official gatekeepers into the social organization ' of the Navy , and then to the ship s Navigation Department . This journey closely parallels the journey through physical space because space is so often used as an element of social organization . As the spatial journey took me to regions with narrower and narrower boundaries , so the social journey leads us through successively narrower levels of social organization . The third journey is a movement through conceptual space, from the world of everyday spatial cognition into the technical world of navigation . This third journey does not really begin until I near the end of the other two .
TI I' ouah Ile MailGate A crisp salute from a young marine in dress uniform at the main ' " " gate s guard shack marked the transition from the street to the " " - base from the civilian realm to the military . The base is a place - of close cropped haircuts and close-cropped lawns . Here nature and the human form are control led , arranged, disciplined , ready to ' " make a good impression . In boot camp inductee s credo is: If it Welcome Aboard 7
' ' moves, salute it . If it doesn t move , pick it up . If you can t pick it " up , paint it white . The same mindset imposes an orderliness and a predictability on both the physical space and the social world of the military base. As a civilian employee of the Navy , I was encouraged to occasionally ride a ship in order to better understand the nature of the " " operational world . But being encouraged by my own organization to ride a ship and being welcomed by the crew are two different things . From the perspectives of the people running a ship , there may be little to gain from permit ting a civilian on board . Civilians , who are often ignorant of shipboard conventions , may require some tending to keep them out of trouble . They take up living space, which on many ships is at a premium , and if they do not have appropriate security clearances they may have to be escorted at all times .
The. The Palau is an amphibious helicopter nansport . Its warfare mission is to nansport marines across the seas and then deliver them to the battlefields in the 25 helicopters that are carried on board . The helicopters also bring noops back to the ship , which has a small hospital and a complete operating theater . Ships of this class are often mistaken for true aircraft carriers of the sort that carry jet planes. As is the case with true aircraft carriers , the hull is capped by a large flat flight deck which creates an overhang on all sides of the ship . But this flight deck is only 592 feet long , just over half the length of a carrier deck and much too small to handle fixed -wing jets.. About halfway between the bow and the stem , jutting up out of the smooth expanse of the flight deck on the starboard rail , stands a four -story structure called the island . The island occupies the rightmost 20 feet of the flight deck, which is about 100 feet wide . The ship extends 28 feet below the surface of the water and weighs 17,000 tons empty . It is pushed through the water by a single propeller driven by a 22,000-horsepower steam turbine engine. ' Originally , the ships of the Palau s class were planned to have been almost 200 feet longer and to have two propulsion plants and two propellers . However , budget cuts in the early 1960s led to a hasty redesign . In the original design, the off-center weight of the steel island was to be balanced by the second propulsion plant . Chapter1 8
Unfortunately , the redesign failed to take into account the decrease in righting moment caused by the deletion of the second engine. When the hull that is now the Palau was launched , it capsized ! It was refloated , and the steel island was replaced with an aluminum one. The ship was renamed and put into service . The aluminum island is attached to the steel deck with steel bolts . In a wet and salty environment , this forms an electrolyte that causes corrosion of the attachment points between the island and the deck. There is a standing joke among those who work in the island that someday, in a big beam swell , the ship will roll to starboard and the island will simply topple off the deck into the sea. Two levels above the flight deck in the island is the navigation bridge . Also in the island are the air operations office, from which the helicopters are control led , and a flag bridge where an admiral and his staff can work . The top of the island bristles with radar antennae.
TheGator Navy - lie OilerNavIes When I first went aboard the Palau it was tied up at pier 4 with several other amphibious ships . A frigate and a destroyer were tied up to an adjacent pier , but they are part of another navy within the Navy . Membership in these navies is an important component of naval identity . Troop transport is not considered a glamorous job in the Navy . The Palau is part of what is called the amphibious fleet , the portion of the fleet that delivers marines to battlegrounds on land . The amphibious fleet is also known somewhat derogatorily as the " " gator navy . The nickname is apparently derived from a reference to that amphibious reptile , the alligator . While the alligator is not a prototypical amphibian , its aggressiveness may be important " " " " in Navy culture ; salamander navy or frog navy might be too disparaging . " " The aviation community (the airdales ) claims to be the highest -status branch of the Navy . Most others would say that the submarine " " fleet (the nukes ) comes next , although the submariners consider themselves a breed apart. (They have a saying that there are only two kinds of ships in the navy : submarines and targets.) " " Then comes the surface fleet (the black shoes ). Within each of these groups are subgroupings , which are also ranked . In the sur- WelcomeAboard 8
face fleet the ranking descends from surface combatants (cruisers , destroyers , and frigates) to aircraft carriers , then the amphibious fleet , and finally tenders and supply ships . While from the civilian point of view a sailor may be a sailor , in the Navy these distinctions mark important subcultural identities . The perceived differences are based on many factors, including the " " glamor of the expected mission , the sophistication of the equipment , the destructive potential , the stringency of requirements for entry into each area, the quality and extent of the training provided to the members of each community , and the general sense of the quality of the people involved . For a surface warfare officer who hopes to make a career out of the Navy and rise to a high rank , it is not good to be assigned to an amphibious ship for too long . Ships that carry aircraft and air crewmen present a special situation with respect to these groups . Because they have aircraft they have members of the aviation community aboard, but because they are ships they must have members of the surface community aboard. The commanding officer of an aircraft carrier is always a - member of the air community a measure of the notion in the navy ' that the air wing is the raison d efre of a ship that carries aircraft . The friction between the air community and the surface community may be manifested in subtle and not -so-subtle ways . If members of the air community account for the majority of the high - ranking positions on a ship , junior surface warfare officers may " " complain that junior airdales are given more opportunities for qualification and advancement . An amphibious transport with an air wing is an even more complicated situation . Here members of the surface and air groups interact . And when marines are aboard an amphibious ship , there is also sometimes friction between the sailors and the marines . These patterns of differentiation are present at all levels of organization in the military , from the broadest of interservice rivalries to distinctions between the occupants of adjacent spaces on the ship . Such effects are present to some degree in many social organizations , but they are highly elaborated in the military . Much of the establishment of identity is expressed in propositions like this : " ' We are the fighting X s. We are proud of what we are and what we " " do. We are unlike any other group . The unspoken inference is If " you do something else, you cannot be quite as good as we are. Identities are also signaled by insignia and emblems of various Chapter1 10
kinds . In the officer ranks , breast insignia denote which navy one is in . Aviators wear wings , submariners wear dolphins , surface warfare officers wear cutlass es. Within each part of the surface fleet , there are strong identities associated with specific ships . Ships have stirring nationalistic or patriotic mottoes , which are often inscribed on plaques, baseball - caps, t shirts , and coffee mugs. Many ships produce yearbooks. The bond among shipmates is strongest when they are off ship . There is ' less of an identification with the class of one s ship , but some classes of ship are considered more advanced (less obsolete) and more glamorous than others. The military institutionalizes competition at all levels of organization . Individuals compete with one another , and teams of individuals are pitted against other teams. Ships compete in exercises , and branch es of the military compete for funding and the opportunity to participate in combat . Aboard a ship this competitiveness " manifests itself in a general opinion that we in our space know what we are doing , but the people just on the other side " of the bulkhead do not . These sentiments can arise in situations where the successful completion of some task relies on cooperation between individuals in different spaces. Sometimes the larger system may fail for reasons having to do with the interactions of the units rather than with any particular unit ; still , each unit needs to attach blame somewhere , and the alleged incompetence of some other unit is the easiest and most understandable explanation .
Acr O8811e Brow A sailor standing outside a guard shack glances at the identification badge of each person passing onto the pier . Walking onto a pier ' between two ships of the Palau s class is like walking into a deep canyon with overhanging gray walls and a dirty concrete floor . The canyon is vaguely threatening . It is noisy , and the hulls of the ships seem to box in the whine of motors and the hiss of compressed air . There are trucks and cranes on the pier , and cables are suewn across the pier and suspended in space over the narrow band of greenish water between the pier and the hulls . Floating in the water between each ship and the pier are several crude rafts called " " camels and a work barge. The camels keep the hull of the ship far enough away from the pier so that the broad flight deck flaring out at the top of the hull does not overhang the pier . WelcomeAboard 11
To board the Palau , I climbed a sort of scaffold up a few flights of gray metal stairs to a gangplank (in Navy parlance , the brow) that reached from the top of the scaffold to a huge hole in the side of the ship . The hole was at the level of the hangar deck (also called the main deck), still several levels below the flight deck. At the top of the brow was a security desk where the officer of the deck (ODD) checked the identification cards of sailors departing from and returning to the ship . Sailors stepping aboard turned to face the stem ' of the ship , came to attention and saluted the ship s ensign (flag), which flew on a staff over the fantail and was thus not visible from the brow . Before visiting the ship , I had been given the NPRDC Fleet visi - ' tor s guide of basic information , which included the following instructions " for proper performance of the boarding ritual : At the top of the brow or accommodation ladder , face aft toward the colors (national ensign) and pause at attention . Then turn to the ODD, ' pause briefly at attention , and say, Request permission to come ' aboard, Sir . State your name, where you are from , the purpose of " your visit and the person you wish to see. This little ritual is a symbolic pledge of allegiance to the ship before boarding . Visitors to the ship wait in limbo at the security desk, neither ashore nor officially aboard, while word of their arrival is sent to their onboard host. The actual permission to go aboard must have been arranged in advance. ' The ship s official gatekeeper is normally the executive officer (abbreviated XO). The commanding officer , the executive officer , and the department heads form the primary administrative structure of the ship . Every ship in the Navy is organized into a number of departments . Each department is supervised by an officer . In large departments , the department head may supervise less senior officers , who in turn supervise the enlisted personnel who do virtually all the actual work on the ship . Before embarking , I was required to convince the XO that I had something to offer the navy and that I would not cause undue aggravation while aboard. In a brief and somewhat discouraging interview with the XO, it was agreed that if the navigator was willing to tolerate my presence in his department , I could come aboard and work with the navigation team. After getting past the XO , I made a date to have lunch with the ' navigator . I met him in the officer s dining area (the wardroom ), and during our discussion we discovered a shared past. While a cadet at Chapter1 12
the Naval Academy , the navigator had served as racing tactician aboard a particular racing sloop that had been donated to the academy. The sloop was subsequently sold to a friend of mine , and I had sailed aboard it as navigator and racing tactician for 8 years. The discovery of this extraordinary coincidence helped cement our ' friendship and secured the navigator s permission for my work aboard the Palau . With my prearranged permission to sail , and ' with the navigator s blessing , I waited at the security desk. An escort at the security desk and led me through the huge dark cavern of the hangar deck. We detoured around several parked helicopters and skirted forklifts and pallets of materials . We ducked through a hatch in the wall of the hangar deck and began the climb up a series of narrow steep ladders to the navigation bridge . (On a ship , tIoors are called decks, walls are called bulkheads or partitions , corridors are called passageways, ceilings are called overheads, and stairs are called ladders .)
R8 CO il Ci I1glie Ch8t- lie World Navigation is a collection of techniques for answering a small " number of questions, perhaps the most central of which is Where am 11" ' ' What does the word where mean in this question ? When we say or understand or think where we are, we do so in terms of some " " representation of possible positions . Where am I? is a question about correspondences between the surrounding world and some representation of that world . Where am I right now as I write this ? I am at my desk, in my study . The window in front of me faces the garden; the door over there leads to the hallway that leads to the remainder of the house. ' My house is on the Pacific coast, north of the university . I m on the ' western edge of the North American continent . I m on the planet Earth circling a minor star in the outer portion of an arm of a spiral galaxy . In every one of these descriptions , there is a representation of space assumed. Each of these descriptions of my location has meaning only by virtue of the relationships between the location described and other locations in the representation of space implied by the description . This is an absolutely fundamental problem that must be solved by all mobile organisms . Whether the map is internal or external , whether it is a mental image of surrounding space (on whatever scale and in whatever Welcome Aboard 13
terms) or a symbolic description of the space on a piece of paper , I must establish the correspondence of map and territory in order to " " answer the question Where am I? One of the most exciting moments in navigation is making a landfall on an unfamiliar coast. If I am making a landfall on a high island or a mountainous coast, as I approach the land , I first see just the tops of mountains , then I see the lower slopes, then the hills , and finally the features on the shoreline itself . Now , where am I? Turning to my chart , I see that I had hoped to meet the coast just to the south of a major headland . Perhaps that big hill I can see across the water on the left is that headland . And perhaps that high peak off in the haze, inland , is this peak shown on the chart . Hmm , according to the ~hart it is only supposed to be meters high . It seems far away and higher than that . Perhaps it is something else, something too far inland to be printed on the chart . Through considerations like these, a navigator attempts to establish a coherent set of correspondences between what is visible in the world and what is depicted on a chart . Some charts even provide small profiles showing the appearance of prominent landmarks - from particular sea level vantage points . The same sort of task confronts any of us when , for example , we walk out of the back door of a theater onto an unfamiliar street. Which way am I facing ? Where am I? The question is answered by establishing cor- respondences between the features of the environment and the features of some representation of that environment . When the navigator is satisfied that he has arrived at a coherent set of corre- " spondences, he might look to the chart and say Ah , yes; I am here, " off this point of land . Now the navigator knows where he is. And it is in this sense that most of us feel we know where we are. We feel that we have achieved a reconciliation between the features we see in our world and a representation of that world . Things are not out of place. They are where we expect them to be. But now suppose " someone asks a navigator How far are we from the town at the " head of that bay? To answer that question , simply having a good sense of the correspondences between what one sees and what is depicted on some representation of the local space is not enough. Now more precision is required . To answer that question the navigator needs to have a more exact determination of where he is. In particular , he needs to have a sense of his location on a representation of space in a form that will permit him to compute the answer to the question . This is position fixing . It is what one does Chapter 1 14
when just having a sense of reconciliation between the territory and the map is not enough .
UpIII Ladder From the hangar deck the escort led the way up three steep ladders in a narrow stairwell filled with fluorescent light , stale air , and the clang of hard shoes on metal steps. The decks of a ship are numbered starting with the main deck. On most ships , the main deck is " " defined as the uppermost deck that runs the length of the ship . On ships that have a flight deck above a hangar deck (this includes aircraft carriers and amphibious helicopter transports such as the Palau ) the hangar deck is the main deck. Immediately below the main deck is the second deck, and below that the third deck, and so on down to the hold . Above the main deck, the decks are designated " " levels and are numbered 01, 02, . . . , increasing in number with altitude . We stopped periodically on deck platforms to allow sailors going down to pass. Foot traffic on ships generally moves up and forward on the starboard side and down and aft on the port side. However , the layout of the hangar deck limits the number and location of ladders , and in order to shorten the route my escort was taking me against the traffic . We climbed into a small busy foyer , and through an open hatch I caught a breath of fresh air and a glimpse of the flight deck in the sun. Men in overalls were working on the hot , rough black surface. We continued upward , now climbing inside the narrow island . One ladder pitch above the flight deck we came to the 04 level . The door leading to the flag bridge , where an admiral and his staff would work , was chained and padlocked . One more ladder brought us to the 05 level .
MIt8YI1d6I ~ ui& The men and women in the military are divided into two broad social classes: officer and enlisted . An officer must have a college degree and is commissioned (authorized to act in command ). In the Navy , members of both classes believe in the reality of differences between officers and enlisted personnel . The lowest -ranking officer is superior in the command structure to the highest -ranking enlisted person . The distinction between officers and enlisted is marked by uniforms , by insignia , and by a complex set of rituals . The simplest of these rituals is the salute, of course, but the Welcome Aboard 15
courtesies to be extended by enlisted to officers include clearing a passagewayon the approach of an officer and refraining from overtaking an officer on foot until permission has been granted.
E* t8d Rates- R8ti1g8 Enlisted personnel are classified according to pay grade (called rate) and technical specialization (called rating ). As Bearden and " Wedertz (1978) explain : A rating is a Navy job- a duty calling for certain skills and attitudes . The rating of engineman , for example , calls for persons who are good with their hands and are mechanically inclined . A paygrade (such as E-4, E-5, E-6) within a rating is called a rate. Thus an engineman third class (EN3) would have a rating of engineman , and a rate of third class petty officer . The term petty officer (PO) applies to anyone in paygrades E-4 through E-9. " E-ls through E-3s are called non -rated personnel . The enlisted naval career begins with what is basically asocial - ization period in which the recruit is indoctrinated into basic military policy and acquires the fundamental skills of a sailor . The rates through which a recruit passes in this phase are seaman recruit , seaman apprentice , and able-bodied seaman. Once socialized , a seaman learns the skills of a particular job specialization or rating . An enlisted person is considered a real member of a rating when he becomes a petty officer (see below ). The enlisted personnel in the Navigation Department are members of the quartermaster rating . ' They have an insignia (a ship s wheel ) and an identity distinct from other ratings . They are generally considered to be relatively intelligent , although not as smart as data processing specialists . For enlisted personnel , rating insignia denote occupational fields . A petty officers is not a kind of commissioned officer (the type of ' ' ' officer referred to by the unmarked term officer ); the label petty ' officer simply designates an enlisted person who is a practicing members of some rating . There are two major levels of petty officer , with three rates within each. One moves through the lowest of these levels while learning the skills of the speciality of the rating . One advances through petty officer third class, petty officer second class, and petty officer first class. A petty officer third class is a novice in the speciality ,and may perform low -level activities in concert with others or more autonomous functions " under instruction " . A petty officer first class is expected to be fully competent in the rating . Chapter 1 18
The next step up in rank moves one to the higher of the enlisted rates and is usually the most important transition of an enlisted ' person s career. This is the move to chief petty officer (CPO). This change in status is marked by a ritual of initiation which is shrouded in secrecy. Just what happens at a chiefs initiation is supposed to be known only by chiefs . However , much of what happens apparently makes for such good story telling that it cannot " " be kept entirely in confidence . It is common knowledge that these initiations frequently include hazing of the initiate , drunkenness , and acts of special license . Making chief means more than getting a bigger pay packet or supervising more people . Chiefs have their own berthing spaces (more private that general enlisted berthing ) and their own mess (eating facility ). On many ships the chiefs mess is reputed to be better than that of the officers. Chiefs are also important because they are the primary interface between officers and enlisted personnel . Since they typically have from 12 to 20 years of experience in their speciality , they often take part in problem -solving sessions with the officers who are their supervisors . Some chief petty officers have a considerable amount of autonomy on account of their expertise (or, perhaps , their expertise relative to the supervising officer .) Chiefs frequently talk about " " having to break in a new officer , by which they mean getting a supervising officer accustomed to the fact that the chief knows more than the officer does and is actually in charge of the space and the people in it . Officers who directly supervise lower -level enlisted personnel risk undermining the chain of command and incurring the resentment of a chief who feels that his authority has been usurped . Once one has made chief , there are still higher enlisted rates to be attained . After approximately 20 years of service a competent person may make senior chief , and after perhaps 25 years of service (being now of about the same age as a captain ) one may make master chief . That is normally the end of the line for an enlisted person . There are, however , some ranks that fall between enlisted and officer . A chief may elect to become a chief warrant officer or a limited duty officer (LDO). A chief who becomes an LDO is commissioned as an ensign and may begin to rise through the officer ranks . Few chiefs take this path . As one senior " chief asked rhetorically , Why would I want to go from the top of one career to the bottom of another ?" While an enlistee may have preferences for certain ratings , the choice of a rating is not entirely up to the enlistee . Aptitude -test WelcomeAboard 17
scores are also used to place people in various specialities . The fact that people are screened contributes to widely held stereotypes concerning the intelligence of those in various ratings . For example ' , boiler technicians (BTs) and machinist s mates (M Ms), who ' run a ship s propulsion plant and who may go weeks without seeing the light of day, are often the butt of jokes about their low intelligence . Data processing specialists , on the other hand , are generally thought to be bright . The ship , as a microcosm , manifests the same patterns of competing identities that are seen among the specialties in the Navy as a whole . From the point of view of the bridge personnel there may be little apparent difference between ' machinist s mates and boiler technicians , but down in the propulsion spaces the perceived differences are many . Machinist mates " " call boiler technicians bilge divers , while boiler technicians call ' " " machinist s mates flange heads. Mostly , this is good-natured ' teasing; name calling is a way of asserting one s own identity . At all levels of organization we see attempts to establish identity by distinguishing oneself from the other groups . This is relevant to the discussion that follows because the dynamics of the relationships among the people engaged in the task of navigation are in part constrained by these identities .
0IIcer R81ks Military officers are managers of personnel and resources. In general , their job is not to get their hands dirty , but to ensure that those who do get their hands dirty are doing the right things . Unlikeen - listed persons, officers do not have narrowly defined specialities . ~ officer pursues a career in one of the broad areas described above: air , surface, or submarine warfare . Within that area, there are sub special ties such as engineering and tactics . Officers are initially commissioned as ensigns. Ensigns have a tough lot . They are more visible than the lowest enlisted rates, and ' " they certainly are given more responsibility , but often a fresh- " caught ensign knows little more about the world of the ship than the seaman recruit .
FldIg One', WayAro I I Ida Sh ~ A ship is a complicated warren of passages and compartments . Every frame and compartment is numbered with a code that Chapter1 18
indicates which deck it is on, whether it is to port or starboard of the centerline , and where it is in the progression from stem to stem . Navigating inside a ship can be quite confusing to a newcomer . Inside the ship , the cardinal directions are forward and aft, port and starboard , topside and below , and inboard and outboard ; north , south , east and west are irrelevant . On large ships , orientation can ' be a serious problem . In the early 1980 s the Navy sponsored a research project to work on wayfinding in ships . The ship is composed of a number of neighborhoods . Some are workplaces , some are residential . Some are officially dedicated to recreation , others are unofficially recreational . The fantail on some ' classes of ships , for example , is a place to hang out . Officers accommodations and eating facilities are in a section of the ship " " called officer country . The chief petty officers have a similar area, " " called CPO country . Enlisted personnel are supposed to enter these areas only when they are on official business. They are supposed to remove their hats when entering any compartment in these neighborhoods . Some passageways inside the ship are major thoroughfares ; others are alleys or culs -de-sac. A visitor quickly learns to search out alternative pathways , because corridors are frequently closed for cleaning or maintenance .
On" ~ Leve I As my escort and I arrived at a small platform on the 05 level , to the - - right was a floor to ceiling partition painted flat black . Behind the partition stood an exterior doorway that led out to the starboard " " wing bridge . The partition forms a light trap that prevents light from leaking out at night when the ship is running dark . To the left was a dark corridor that led to a similar doorway on the port side of the island . Above us, the ladder continued upward one more level to the signal bridge . Ahead lay a narrow passageway. Forward along the left side of the passageway were two doors. Behind the ' - first was the captain s at sea cabin . He has a nicely appointed quarters below , but he takes meals and sleeps in this cabin during operations that require him to stay near the bridge . The next door opened on the charthouse . At the end of the passageway, about 25 feet away, was a door that led to the navigation bridge or pilot - house. The charthouse is headquarters for the Navigation Department . This small room , crowded with navigation equipment , two desks, a Welcome Aboard 18
safe, and a chart table , enjoys a luxury shared by only a few spaces on the ship : a single porthole through which natural light may enter and mix with light from the fluorescent lamps overhead . The charthouse is one of several spaces under the control of the Navigation Department . Navigation personnel not only work in these spaces, they are also responsible for keeping them clean . Since the ' bridge is one of the main work areas of the ship s captain , it is thought to be especially important to keep it looking nice . While in port , Navigation personnel polish the brass on the bridge . Because ' the captain s at-sea cabin is adjacent to the charthouse , members of the Navigation Department tend to work more quietly there than they might in other parts of the ship . Since the average age of a sailor is under 20 years, a certain amount of playful horsing around is expected in many parts of the ship , but is not tolerated on the 05 level . The Navigation Department is responsible for all of the spaces on ' the 05 level with the exception of the captain s at-sea cabin . It is also responsible for the secondary or auxiliary conning station " " ( Secondary Conn )- a completely redundant navigation bridge located in the bow , just under the forward edge of the flight deck. ' Secondary Conn is manned by the ship s executive officer and a complete navigation team whenever the ship is at general quarters (battle stations ). This is done because the primary navigation bridge in the island is very vulnerable if the ship comes under attack. Modem anti -ship missiles home in on electromagnetic radiation . Because the radar antennae on the top of the island are the principal sources of such radiation on the ship , the island is the most likely part to be hit by a missile . If the primary navigation bridge is destroyed , the ship can be control led from Secondary Conn under the command of the executive officer . Secondary Conn is a space assigned to the Navigation Department and is a duty station for Navigation personnel , but it will be of little interest to us with regard ' to the normal practice of navigation . The ship s extensive library of charts and navigation forms is stored in this space. The Navigation Department is supervised by the Navigator . At ' the time the observations reported here were made, the Palau s Navigation Department consisted of the Navigator and seven enlisted " " men. The title Navigator refers to the position as head of ' the Navigation Department rather than to the officer s technical speciality . Though it is expected that an officer who serves as Navigator aboard any ship will know enough about navigation to Chapter1 .
supervise the working of the Navigation Department , Navigators seldom do any navigating themselves . The work of the Navigation Department is carried out by enlisted personnel of the quartermaster rating under the direction of the Assistant Navigator (a quartermaster chief ).
NavigatingLaIrge . . While a naval vessel is underway , a plot of its past and projected movements is maintained at all times . Such complete records are not always kept aboard merchant vessels and are not absolutely essential to the task of navigating a ship in resbicted waters. It is " " possible for an experienced pilot to eyeball the passage and make judgements concerning control of the ship without the support of the computations that are carried out on the chart . Aboard naval vessels, however , such records are always kept - primarily for reasons of safety, but also for purposes of accountability . Should there be a problem , the crew will be able to show exactly where the ship was and what it was doing at the time of the mishap . Day and night , whenever a ship is neither tied to a pier nor at anchor , navigation computations are performed as frequently as is required to ensure safe navigation . During a long passage, navigation activities may be performed almost continuously for weeks or even months on end. Most of the time the work of navigation is conducted by one person working alone. However , when a ship leaves or enters port , or operates in any other environment where maneuverability is re- sbicted , the computational requirements of the task may exceed the capabilities of any individual ; then the navigation duties are carried out by a team. The conning oJ Jicer is nominally responsible for the decisions about the motion of the ship , but for the most part he does not make the actual decisions . Usually , such decisions are made by the Navigation Department and passed to the conning officer as recommendations " , such as Recommend coming right to 0 1 7 at this " time . The conning officer considers the recommendation in the ' light of the ship s overall situation . If the recommendation is appropriate , he will act upon it by giving orders to the helmsman , who steers the ship , or to the leehelmsman , who controls the engines . At all times when the ship may have need of navigational information , someone from the Navigation Department is at work and ready to do whatever is required . The navigation team per- WelcomeAboard 21
forms in a variety of configurations , with as few as one and as many as six members of the Navigation Department working together . In every configuration there is one individual , designated the quartermaster of the watch , who is responsible for the quality of the ' work performed and who serves as the department s official interface with other departments aboard ship . Navigation is a specialized task which , in its ordinary operation , - confronts a limited set of problems , each of which has a well understood structure . The problem that confronts a navigator is usually not one of figuring out how to process the information in order to get an answer ; that has already been worked out . The problem , in most instances , is simply to use the existing tools and techniques to process the information gathered by the system and ' to produce an appropriate evaluation of the ship s situation or an appropriate recommendation about how the ship should proceed in order to get where it is supposed to go. The navigation activity is event-driven in the sense that the navigation team must keep pace with the movements of the ship . In contrast with many other decision -making settings , when something goes wrong aboard a ship , it is not an option to quit the task, to set it aside momenta rily , or to start over from scratch . The work must go on. In fact, the conditions under which the task is most difficult are usually the conditions under which its correct and timely performance is most important .
.. ..I 1118A. -&.wl... . ldenaty Having said something about how naval personnel establish their own identities , I should also say something about how they and I negotiated an identity for me. In the course of this work I made firsthand observations of navigation practice at sea aboard two aircraft carriers (the Constellation and the Ranger) and two ships of the amphibious fleet (the one known here as the Palau and the Denver). Aboard the aircraft carriers , I worked both on the navigation bridge and in the combat information center . I made a passage from San Diego to Seattle, with several stops, aboard the Denver. I also interviewed members of the Navigation Departments of five other ships (the Enterprise , the Beleau Wood, the Carl Vinson , the Cook, and the Berkeley) and had a number of informal conversations with other navigation personnel . Chapter1 22
The events reported here come mainly from operations in the Southern California Operations (So Cal Ops) area aboard the Palau . I also worked with the crew while the ship was in port . I logged a total of 11 days at sea over a period of 4 months . First came a week- long hip during which I observed the team, got the members used to my presence, and got to know them . During this trip , I only took notes and made a few still photos and audio tape recordings of navigation tasks and interviews with crewmen . On a later hip , I mounted a video camera with a wide -angle lens in the overhead above the chart table in the pilothouse . I placed a stereo tape recorder on the chart table , with one channel capturing the ambient noise and conversation of the pilothouse . The other channel I wired into the sound -powered phone circuit . Because the chief was both plotting positions and supervising the work of the navigation team, I wanted to be sure to capture what he said. I therefore wired him with a remote transmitter and a lavaliere microphone . I used this signal to feed the audio track on the video recording . Thus , I had one video track and three audio tracks to work with . During my time at sea, I took a normal watch rotation . I appeared on the bridge on one occasion or another during every watch period , including the one from midnight to 4 aim. I was accorded privileges appropriate to the military equivalent of my civilian Government Service rank : lieutenant commander . I was assigned a " " ' cabin in officer country , took my meals in the officer s mess, and spent my waking off-watch time either in the charthouse with the navigation crew or in the wardroom with officers. As to what they thought of me, one must begin with the understanding that for military folk the military /civilian distinction stands just below the friend /foe distinction as an element of the establishment of identity . A civilian aboard a ship is an outsider by definition . It was important that the navigator treated me as acol - league and friend , and that the captain normally addressed me as Doctor when we met. Many of the members of the navigation team ' were also aware that I had lunched at least once in the captain s quarters , an honor reserved for visiting VI Ps. Some evidence of what the crew thought of me is available in the video record . Early on, a number of nervous jokes were made on camera about the dangerous potential of the videotaping . In the first 5 minutes of videotaping with this crew , the assistant navigator " told the navigator Everything you say around me is getting recorded " for history , for your court -martial . Welcome Aboard 23
On more than one occasion while he was away from the chart table , the chief of the navigation team explained my work to other members. He apparently forgot that he was being recorded . I discovered these comments weeks later while doing transcription . During my second at sea period , the chief went into the charthouse to check on the fathometer . The fathometer operator asked who I was. The conversation proceeded as follows : ' ' Chief: He s studying navigation on big ships . He s the guy , he makes computer programs for teaching stuff . Like they got a big computer program thing they use in ASW school to teach maneuvering ' boards . It s all computerized . He is the one that makes it . He ' - is the one who makes things like that . He s a psychologist and an ' thropologist . Works for the navy . He s a PhiD . Makes all kinds of strange things . Falhometeroperator : He makes all kinds of strange money too. ' Chief: Yeah, does he? He knows what he is doing . He s swift . He ' just sits and watches and records everything you re doing . Then he puts it all in data, then he starts putting it in a program . Figuring ' out what to do, I don t know . - My most intensive data collection was carried out on a four day exercise during which the Palau left port , steamed around the operations area for two days, reentered port , and anchored in the harbor overnight . The next morning the ship left port again for another day of exercises. Finally , it entered port again and returned to its berth at the 32nd Street Naval Station . It was during the last entry to port that the crisis reported in the opening pages of this book - occurred . The quality of the recording from the sound powered phone circuit was poor until I discovered a better way to capture the signal on the last entry to port . The two entries to and exits from port were recorded from the time Sea and Anchor Detail was set until the navigation team stood down . This procedure produced video and audio tape recordings of about 8 hours of team activity . Additional recordings were made at various times during Standard Steaming Watch . In addition to the video and audio records , I took notes during these events of any aspects of the situation that I noticed that could not be fully captured on the tapes. Even with the wide -angle lens, the video camera captured only the surface of the chart table . This permit ted me to identify features on the chart and even to know which buttons of a calculator were pressed, but it Chapter1 24
meant that many events of interest were not captured on tape because they occurred out of camera range. Transcribing the tape recordings was a very difficult process. At times there were four or more conversations happening simultaneously in the pilothouse . To make matters worse , ships are noisy places. There are many kinds of equipment on the bridge that create ' background noises. The bosun s mate pipes various announcements from a station just aft and inboard of the chart table , and his whistle blowing and his public -address messagessometimes drown out all other sounds. Helicopters may be operating on the flight deck or in the air just outside the pilothouse . It was often necessary to listen to each of the three audio tracks separately in order to reconstruct what was being said, and still in many cases the full content of the tapes cannot be deciphered . Because of the placement of the microphones , however , the coverage of the verbal behavior of the members of the navigation team was uniformly good. Only rarely was it impossible to determine what was being said with respect to the navigation task. I did much of the transcription myself , for three reasons. First , this is a technical domain with many specialized words in it . We know that hearing is itself a constructive process and that ambiguous inputs are often unconsciously reconstructed and cleaned up on the basis of context . Lacking context , other transcribers could not hear what I could hear in the tapes. For example , an untrained transcriber without expectations about what might be said during " " " an anchoring detail transcribed thirty fathoms on deck as thirty " phantoms on deck. Navigationese is a foreign language to most people , and quality transcription cannot be expected from a transcriber who is not fluent in it . Second, since there were many speakers, the fact that I knew them personally helped me distinguish the identity of speakers where it was not clearly evident from the content of a statement who was speaking. Third , and most important , there is no better way to learn what is actually in a recording than to listen to it the many times that one must in order to produce a good transcription . (Over a period of about a year, one transcription assistant did develop enough familiarity with the subject to provide usable transcriptions .) The fact that listening is reconstructive introduces the possibility of distortions in the data driven by my expectations . I will attempt to deal with that by making the ethnographic grounds for my interpretations explicit . Welcome Aboard .
In the pilothouse I tried not to participate , but only observe. On only one occasion did I intervene , and that was a case in which I felt that by failing to speak I would put a number of people in serious danger. My intervention was a brief sotto voce comment to the navigator , who resolved the situation without indicating my role in it . It was clear that I knew more about the theory of navigation than the members of the crew I was studying with the exception of the ' ship s navigator and the quartermaster chief . Of course, knowing the theory and knowing the nature of the practice in a particular setting are two quite different things . In no case did I know more ' about an individual s relation to the practice of navigation than that individual . Still , this is an unusual situation for an ethnographer . The web of constraints provided by cultural practices is important both to the people doing the task and to the researcher. For the performers , it means that the universe of possible activities is - closely bounded by the constraints . For the researcher, the activ ities that are observed are interpreted in terms of their reflection of the constraints . My many years of studying and practicing navigation made me a particular sort of instrument , one in which the constraints of the domain were present. My interpretations of the actions of the members of the navigation team were informed by many of the same constraints that were guiding their behaviors . But there was more . Because I attempted to continually make these constraints explicit , and to conceive of them in a computational sense as well as in the operational sense required of the navigation team, my interpretations were not simply those of a native . A few months of field work is, for an anthropologist , a rather a short visit . Many aspects of the military culture go unreported here because I am not confident about their organization and meaning on the basis of such a short exposure . I did have 5 years of employment as a civilian scientist working for the Navy , and that gave me many opportunities to observe aspects of military organization . The coverage of navigation practice is adequate, I think , because of the opportunity on my second at-sea period to videotape the navigation operations on the bridge . How different would the story be if the observations had been made aboard another ship ? I do not believe that the culture would permit it to be very different . The information processed by the navigation team may move more or less efficiently , and the individual quartermasters may have better or poorer relationships with Chapter1 .
one another , but the tasks remain , and the means of performing the tasks are standardized throughout the fleet. The crews of different ships may meet the requirements of navigation more or less capably , but they must nevertheless solve these particular tasks in the limited number of ways possible . In fact, I made observations aboard several ships , and my colleague , Colleen Siefert , did so on yet another ship . The differences we observed across ships were minor . The ship Colleen observed had more quartermasters available and was therefore able to organize its navigation team in a slightly different way ; that however , does not present a challenge to my framework or to my basic descriptions of the nature of the cognition at either the individual or the group level .
Onlie ~ : Stand8dSt8 &I7~.gWatch ' At the forward end of the 05 level s passageway is the door to the navigation bridge or pilothouse . It is here that the most important part of the navigation work is done. The pilothouse occupies the forward 18 feet of the 05 level of the island (see figure 1.1). Outward - canting windows extend from chest height to the overhead on both sides and the front of the pilothouse . The windows on the port side and forward overlook the flight deck. All work tables are mounted on substantial bases on a light greenish linoleum floor . The walls , the cabinets , and the equipment stands are thickly coated in light gray paint . The overhead is flat black and tangled with pipes and cables, their identities stenciled on them in white . ' The polished brass of ship s wheel and the controls for the engine- order telegraph stand out in the otherwise drab space. The activities of the Navigation Department revolve around a computational ritual called the fix cycle. The fix cycle has two major parts: determining the present position of the ship and projecting its future position . The fix cycle gathers various bits of information ' about the ship s location in the world and brings them ' together in a representation of the ship s position . The chart is the ' positional consciousness of the ship : the navigation fix is the ship s internal representation of its own location . ' When I first made it known to a ship s navigator that I wanted to know how navigation work was performed , he referred me to the Navigation Department Watch Standing Procedures, a document " ' " that describes the watch configurations . It s all in here, he said.
Welcome Aboard 27
'
21
" " -
I I
'
D s CO
0 0 Rajars
table table
Gyro Gyro
L _ : : : : : :
roI ' Ql
i . ~ : : ~ . J
netic netic
Mag Mag
Capfs Capfs
table table
0
r l ' '
16 ~ ~ ~ : : :
00 00 ~
l ~ : : : : : _ 1
Chart
table
Capfs Capfs
Helm Helm
chair
Pilothouse
~ ~
' '
mate s Bsn s
Contact Contact board
Door
Door Door
table table
Door Door
0 0 0
Pelorus
PelorusFathometer PelorusFathometer
Charthouse Charthouse
Starboard Port Port
P P
assageway assageway
wing wing wing
I I I . I I I I . I I I I . I
Fig In 1.1 A planview of . . pI~ and. . dlar Itouse. The~ - of lie navigationteam do mostof lleir workat lie cIwt mb Ie, on lie wiIgs, andit lie clld I O I J88. Theheavy line f8p1 ' 8s811tslie exteriorIkiI of ' " ship. Upin ' " d~ is forwardon ' " ship.
" You can read this and save yourself the trouble of standing " watch. Of course it is not all in there, but the normative description in the Proceduresis not a bad place to start. It is the Navigation ' " " Departments omcial version of the organization of its work. This document is one of many symbolic forms in which navigators " " represent themselves to themselves and to one another (Geertz 1983). Becausethe procedures refer to objects and places that are part of shipboard navigation culture, understanding these procedures will require us to explore the environment of navigation. While conducting this exploration, we should keep in mind that the Chapter1 .
' descriptions of navigation work that appear in a ship s documents and in various navigation publications must be taken as data rather than analysis . ' In this section I will attempt to use the ship s documents as a guide to the task of navigation . The specifications presented in the Watch Standing Procedures describe actions to be taken and equipment and techniques to be used. First I will present the normative descriptions and try to provide the sort of background information that might be provided by a native of the navigation culture , in the hope that this will make these things meaningful to a reader who is not a practitioner of the art. Later I will present an analysis of the procedures , tools , and techniques that will be grounded in information -processing theory rather than in the world of ship navigation . ' The Palau s normal steaming watch procedures are introduced as follows :
Whilein normalsteaning condition at - . lie followingwatch procedt N wiNbe adhered to as closelyas DO SIible , modifiedas nec8ary by sib Jationsbey OI1d the conboIof thewatch stander.
In normal steaming, a single quartermasteris responsible for all the navigation duties. The procedures described in the document are taken seriously , although it is recognized that it may not be possible to execute them as described in all circumstances . The normative procedures are an ideal that is seldom achieved , or seldom achieved as described .
1118Prinary Duty of lie QMOW When the Navigation Department is providing navigation services to the ship, a particular quartermasteris designatedas the quarter master of the watch (QMOW) at all times. According to the procedures , ThePrinary Duty of lie QMOWis. . safenavigation of. . ship. ToUIis end he shalt (I) Fix. . positionoflie shipby 81 '1n d me UIods av818 )le. (1) ~ fixeswill be plotted . (~ w...I infonnationisavailable , a fixwill be plotted at I Bt everyhour , wtt.Iln openocun bansit. (3) WhenwiUIin VIsual or ~ ~ ofland , afix will be plotted at I88t everyfifteen mll Mites. (I) VIsualbe Iri Igswill take priority . OQ FillIn wiUI ~ . r~:~ . (4) Fixesmay be obtained from 8IY combination ofu . fo Ilowl Ig aIrces : (I) Yi8t8belri Igs Welcome Aboard 8
00 R8iarranges 010 R8Iarb88tngs (Iv) FdIometerOill oflOt I Idings, bottomCC X1tourtng , Mout hopping) M Nav Sat (vi) Omega (vii) Ce I8tia I ob I8vatior. (5) Fix. obtai18dfrom ~ radarsou . - willconsist of . leastII . . LO Ps. ' (b) Projectlie ships trackby deadrlckoring to a sufticientBIgth of tine UIat8IFI d8Iger jiI~ ~ to lie shipfrom land , shoals 011 fixedd81gers , violationof Internationalwaters wli benoticed well In advance oflie shipacb I IIly st8Idilg into d8Iger departinglegaJ /usigned water I.
Items a and b in this document describe the two main parts of the ' fix cycle : fixing the ship s position and projecting its track . The procedures of dead reckoning will be explained in detail in chapter 2. The plotted fix is a residue on the chart of a process that gathers ' and transforms information about the ship s position . A succession of fixes is both a history of the positions of the ship and a history of the workings of the process that produced the position information . The requirement that all fixes be plotted ensures a complete history of positions and provides certain opportunities to detect and correct faults in the process that creates the history . The interval between fixes is set to 60 minutes in open waters and no more than 15 minutes when the ship is in visual or radar contact with land . Near land , the ship may stand into danger more quickly than when in the open ocean. Sailors know that it is not the open ocean that ' sinks ships , it s all that hard stuff around the edges. The increased frequency of fixes near land is intended to ensure that dangers are anticipated and avoided . Visual bearings are given priority because they are the most accurate means of fixing position . The potential sources of position information are listed roughly in order of their accuracy and reliability . The procedure states that fixes may be obtained from any combination of a number of sources. Let us briefly consider the nature of these sources and the kinds of information they contribute to fixing the position of the ship .
So. . cesof Infomlltion for PoI Iaon FIxiIg
VISUAL BEARINGS The simplest way of fixing position, and the one that will concern us most in this book, is by visual bearings. For this one needs a chart of the region around the ship and a way to measure the Chapter1 .
- direction (conventionally with respect to north ) of the line of sight connecting the ship and some landmark on the shore. The direction ' of a landmark from the ship is called the landmark s bearing . Imagine the line of sight in space between the ship and a known landmark . Although we know that one end of the line is at the ' landmark and we know the direction of the line , we can t just draw a line on the chart that corresponds to the line of sight between ' ship and landmark , because we don t know where the other end of the line is. The other end of the line is where the ship is, and that is what we are trying to discover . Suppose we draw a line on the chart starting at the location of the symbol for the landmark on the chart and extend it past where we think the ship is- perhaps off the edge of the chart if we are really ' unsure . We still don t know just where the ship is, but we do know it must have been somewhere on that line when the bearing was observed. Such a line is called a line of position (LOP). If we have another line of position , constructed on the basis of the direction of the line of sight to another known landmark , then we know that the ship is also on that line . If the ship was on both of these lines at the same time , the only place it can have been is where the lines intersect . The intersection of two lines of position uniquely constrains the location from which the observations were made. In practice , a third line of position with respect to another landmark is constructed . The three lines of position form a triangle , and the size of that triangle is an indication of the quality of the position fix . It is ' sometimes said that the navigator s level of anxiety is proportional to the size of the fix triangle . The observations of visual bearings of the landmarks (direction with respect to north ) are made with a special telescopic sighting device called an alidade . The true -north directional reference is provided by a gyrocompass repeater that is mounted under the alidade . A prism in the alidade permits the image of the gyro- ' compass s scale to be superimposed on the view of the landmark . (The view through such a sight is illustrated in figure 1.2.) The gyrocompass repeaters are located on the wings outside the bridge . Each one is mounted on a solid metal stand just tall enough to extend above the chest-high metal railing that bounds the wing . The most direct access to the port wing from the chart table is through a door at the back of the pilothouse just behind the cap- ' tain s chair . In cold weather , the captain of the Palau does not permit traffic through this door . The only other way to get from the
WelcomeAboard 31
FIg&n1 .2 A view" rough81 alld .te. A prtsmInside UIealid8 Ie superimposesUIeIrnag8 of twocomP8 scalesonto wt ~ Isseen Bra UIe Theim . - scaleIs a reIndicates . telescopicsighl gyrocompass' peater, UIeout . scaleIs fastened toUIe ship 8Id bearilgsrelative to UIe ship s head.
port wing position to the chart table is to go aft on the wing to the hatch that leads to the island stairwell and then come forward ' - through the interior passageway past the captain s at sea cabin and the charthouse . This makes it difficult to get bearings sometimes , because it takes a long time to go around the entire 05 level .
RADAR Radar also provides information for position fixing . The radar antenna ' on the ship s mast transmits pulses of radio magnetic energy as it rotates. When the pulse strikes a solid object , the pulse reflects off the object. Some of that reflection may return to the radar antenna that transmitted it . By measuring the time required for the pulse to travel to the object and return , the radar can compute the distance to the object. This distance is called the range of the object . The direction in which the antenna is pointing when the reflected pulse returns gives the bearing of the object. Radar ranges are more accurate than radar bearings , so they are given priority in position plotting . In practice , radar ranges plotted as circles of position are often combined with visual bearings to produce position fixes . The surface search radar displays are located at the front of the pilothouse on the starboard side. Each is equipped with a heavy black rubber glare shield that improves the visibility of the display in high ambient light . This glare shield prevents two or more people from looking at the scope at the same time . The surface search radar also has non -navigational uses. The Chapter1 .
officer of the deck may use the radar to observe and track other ship traffic . For this , a short range is usually desired . The navigation tasks often require a long range, and there is sometimes conflict between the two users of the scopes. It is not difficult to change from one range to another ; however , in order to obtain the required information after changing ranges, the operator may have to wait for a full rotation of the radar antenna at the new range setting .
FA THOMETER The fathometer is a device for measuring the depth of the water under a ship . It emits a pulse of sound and measures the time it takes the sound pulse to bounce off the sea bottom and return to the ship . The time delay is recorded by the movement of a pen across a piece of paper . The sound pulse is emitted when the pen is at the top of the paper . The pen moves down the paper at a constant speed and is brought into contact with the paper when the echo is detected. The distance the pen navels down the paper before making its mark is proportional to the time required for the echo to return , which is in turn proportional to the depth of the water . If the water is deep, the sound will take longer to return , and the pen will have nave led farther down the paper before coming into contact with it . The depth of the water can be read from the scale printed on the paper. Changing the scale of the fathometer to operate in deeper or shallower water is accomplished by changing the speed at which the pen navels . The paper is mounted on a motor drive that moves the paper to the side a small amount just before each pulse . This results in a continuous graphical record of the depth of the water under the ship . The Palausfathometer is located in the charthouse , so the QMOW must leave the bridge to use it .
NAVSAT Satellite navigation systemshave now becomecommonplace . They are easyto use, and they provide high-quality position information. Their major drawback at the time this research was carried out was that with the number of navigation satellites then available the mean interval between fixes was about 90 minutes . After computing ' the ship s position from the reception of satellite signals, the satellite navigation system continuously updates the position of the ship on the basis of inputs from the gyrocompass (for direction ) and Welcome Aboard &1
' the ship s log (for speed). The NavSat system aboard the Palau (located in the charthouse ) was a box , about the size of a small suitcase that continuously displayed a digital readout of the latitude and longitude of the ship . The fact that NavSat systems must update position with dead reckoning during the long wait between fixes puts NavSat near the bottom of the list of sources of information . With the implementation of the Global Positioning System (GPS), continuous satellite fixes are now available ; the need for dead-reckoning updates of position has been eliminated . The military version of GPS is accurate to within less than a meter in three dimensions . The civilian versions are intentionally degraded to a consider ably lower accuracy . GPS will very likely transform the way navigation is done, perhaps rendering most of the procedures described in this book obsolete.
OMEGA Omega measures the phase difference between the arrival of signals from multiple stations . Omega was intended to provide accurate worldwide position -fixing capability . In practice it is unreliable . Whatever the source of the problems , they are perceived to be so serious that the following warning appears in the Watch Standing Manual .
CAUTION: Positionsobtained from Omega are h ~ 1ysuspect , 11118888Ub Itafi ~ byInformation fromanother source . Inrecent years , a numberof cosily8Id 8nban8ing~ ndlngshave ~ directlyattributable to trustingOrnega . ~ dl8tic decisionsare ! !!: to be madeon unsubstantiated Omegafix. wi UI OUt tie explicitper millionof lie navigm.
If this system is considered to be so unreliable that it merits this strongly worded caution in the written procedures , what is it doing on the ship ? I believe the answer involves an interaction of the organization of military research and funding with the development of technology . Omega is a system that not only went into service before all the bugs could be worked out , it has been overtaken by other superior technologies before the bugs could be worked out . Still , it was bought and paid for by the military , and can, on occasion , provide useful navigation information . ' The Palau s Omega is located in the charthouse .
CELESTIAL OBSERVATIONS By measuring the angular distance of a star above the horizon , an observer can determine his distance from the point on the surface Chapter1 34
of the earth that the star is directly above. This point forms the center of a circle of position . In a celestial sight reduction , each observed celestial body defines a circle of position , and the vessel from which the observations were made must be located at the intersections of the circles of position . Celestial observations appear at the bottom of the list of sources of information . When properly performed , celestial observations provide fairly good position information . There are, however , two major drawbacks to celestial observations . First , they can be performed only under certain meteor- ological circumstances . This makes celestial navigation hard to use and hard to teach. Several senior quartermasters have told me that they would like to teach celestial navigation on training missions in the Southern California operations area, but the combination of air pollution and light pollution (which makes the night sky bright , masking all but the brightest stars and obscuring the line of the horizon ) produces very few occasions suitable for it . Second, the procedures are so computationally complex that , even using a spe- cialized calculator , a proficient celestial navigator needs about half an hour to compute a good celestial position fix . Together these factors lead to infrequent practice of this skill . I believe that in the near future the only navigators who will know how to fix position by star sights will be those sailing on cruising yachts who cannot afford a thousand dollars for a Sat Nav system.
DRAI The Dead Reckoning Analyzer Insb" ument ( DRAI) is one of the most interesting navigational devices. A mechanical analog computer , ' it takes input from the ship s speed log and the gyrocompass and, by way of a system of motors , gears, belts , and cams, continuously computes changes in latitude and longitude . The output of the DRAI is expressed in the positions of two dials : one reads latitude and the other longitude . If these dials are set to the current latitude and longitude , the changes computed by the motions of the internal parts of the DRAI will move them so that their readings follow the latitude and longitude of the ship . The crew of the Palau claimed that when , properly cared for , the DRAI is quite accurate and reliable . Older versions of the DRAI , such as the one aboard the Palau , have been around since the 1940s. Newer versions that do the same computations electronically are installed on some of the newer ships . WelcomeAboard .
PIT SWORD AND DUMMY LOG The pit sword is a device that is extended through the hull and into ' the water to measure a ship s actual speed through the water . The pit sword extends several feet outside the hull and measures speed ' by measuring the water s distortion of a magnetic field . The speed signal generated by the pit sword is fed to speed indicators on the bridge and to all the automated instruments that do dead reckoning : the NavSat, the DRAI , and the inertial navigation systems (if present). If the ship is operating in shallow water , the pit sword cannot be extended from the hull . In this case, or if for any other reason the pit sword cannot be used, the dummy log is used. When a ship is neither accelerating nor decelerating , its speed can be estimated fairly accurately from the rate of rotation of the propeller . The dummy log is a device that senses this rate and provides a signal that mimics what the pit sword would produce at the corresponding speed. Both of these devices are remote from the location of the navigation ' team s normal activities . A display of speed through the water is available on the forward port side of the pilothouse , but it is rarely consulted by the navigation team.
CHRONOMETERS Tkee traditional spring -driven clocks are kept in a special box in ' the Palau s charthouse . Readings are recorded daily so that trends in the behavior of these chronometer 's can be noted . These records are maintained while time signals are available on radio so that if time signals should become unavailable the behavior of the clocks will be known . If , for example , the log shows that a particular chronometer loses a second every day, that same rate of change will be assumed until more reliable time sources are restored .
The diversity of the many sources of navigation information and ' the many methods for generating constraints on the ship s position produces an important system property : the fact that positions are determined by combining information from multiple , sometimes independent , sources of information permits the navigation team to check the consistency of the multiple representations with each other . The probability that several, independently derived , representations are in agreement with one another and are in error is much smaller than the probability that anyone representation is in error . Chapter1 .
AtIII Ch8tTable The previous section described the sources of information that the quartermaster of the watch may use while discharging his primary duty : ensuring the safe navigation of the ship . The information provided by these sources converges on the chart table , where positions are plotted and tracks are projected . The Watch Standing Procedures specify additional constraints on the QMOW that bring us to other aspects of the navigation ' team s task setting : 1118dIart table 8Id nions willbe kept free of exhneous material atd tines. Only' " chart(s) in - , ~ ryJ Mlblcations , ' " logsof tie watch, andnecessary writing /PIotti Ig parlpilernaliawill beon ' " dIarttable .
The chart table is mounted against the starboard wall of the pi - lothouse , just under the large outward -canted windows . It is large enough for full -size navigation charts and tools - about 4 by 6 feet. Under the chart table are a number of locking drawers in which charts , publications , and plotting tools are stored. A locking cabinet for binoculars is mounted on the aft edge of the chart table.
NavigationCh8I8 The most important piece of technology in the position -fixing task is the navigation chart . A navigation chart is a specially constructed model of a real geographical space. The ship is somewhere " " in space, and to determine or fix the position of the ship is to find ' the point on the appropriate chart that corresponds to the ship s position in space. The lines of position derived from visual observations , radar bearings , radar ranges, celestial observations , and depth -contour matches are all graphically constructed on the chart . Latitude and longitude positions determined by NavSat, Omega, or Loran are plotted directly on the chart . A fix may be constructed ' from a combination of these types of information . Navigation charts are printed on high -quality paper in color . " " Natural and cultural features are depicted in a complex symbology (see figure 1.3). The Palau keeps an inventory of about 5400 charts depicting ports and coastlines around the world . A comp ~ete set of charts for current operations are kept on the chart table , and a second complete ' set in the table s drawers . The rest of the charts are kept in a chart library in Secondary Conn . WelcomeAboard 37
1.3 A ctBt m . . SudIa ch8t kICk Id8 klfonn Iti C I I8 )out f8IbI8 boll above8Id below ~ Mvigallon - III ~ . ThIscIwt ~ ~ illiii I~ to S8I ~ H8b0r. Chapter 1 .
nil Sec OI Id8 YDuty of lie QMOW According to the Watch Standing Procedures , Thesecondary duty of VIe QMOW IsVIe keepi1g ofVIe logs of VIe watch .
Those who have experience in the merchant fleet often say that it is not necessary to do all the work of piloting in order to get a large " " ship into port . A good ship driver can, after all , eyeball the movement of the ship and get it down the channel without having positions plotted on short intervals . To say that it is possible to guide a ship down a narrow channel without maintaining the piloting record is not to say that it is easier to do it that way . Even if nothing goes wrong , the plotted and projected positions of the ship on the chart are a useful resource to the conning officer , and while it does require a navigation team to do the work of plotting positions and computing turn points , the task of the conning officer is greatly simplified by the advice he receives from the navigation team. If something does go wrong , the work of the navigation team becomes indispensable in two ways . First , depending upon what it ' is that goes wrong , computing the ship s position and track may become essential to the process of figuring out how to keep the ship out of trouble (see chapter 8 for an example ). Second, the records kept by the navigation team- the chart , the deck log, and the bearing log- are all legal documents . If the ship is involved in a mishap , as soon as it is prudent to do so, all these documents are removed from the chart table and locked in the Executive Officer 's safe. This precaution is taken to ensure that they will not be tampered with before they are turned over to a board of inquiry investigating the incident . These records may be needed to protect the navigation team, the captain , the ship , and ultimately the Navy ' from accusations of negligence or incompetence . The Palau s Assistant Navigator offered the following justification : ' You can go into San Diego by eye. But legally , you can t. If you ' haven t matched all the things and something happens , not neces- ' sarily to you , it don t have to. One of those buoys can float loose in the god damn bay and rub up along side you . Boy, you better have everything covered here, because they are going to tzy to hang the captain . They will tzy to hang him . Unless he can prove with data that everything he did was right . Now . . . the merchant ship ' " wouldn t. They would just say, We were in the middle of the channel . The damn thing hit us, and if there is an expense, fine , " charge the company . Welcome Aboard .
Other records are kept as well . There is a separate log for the gyro compass es (with entries made twice daily ), and another for the magnetic compass es. (The DRAI reading is also recorded in the magnetic compass log at the beginning of each watch .) There is yet another log for the shipschronometers . A fathometer log is kept with the fathometer during maneuvers in restricted waters . A log of ' the ship s position is updated daily .
11IeT . . . , Dutyof III QMOW " The tertiary duty of the quartermaster of the watch is to give all " possible aid to the Officer of the Deck in the conduct of his watch . The Officer of the Deck (OOD) is also normally the conning officer , although he may delegate this duty to a Junior Officer of the Deck. The importance of the relationship between the QMOW and the OOD is reflected in the following excerpt from the Watch Standing Procedures :
TheQMOW will not leave lie exceptto takeDRAI 81d Fau ~ rBlngs. 81dcollect Bridge- NavSat81d ~ fix. asr ~:~ j . If heleaves the bridge , hewi " Informlie OOD, 81dwill absent hinse If foras short a periodof tine asP O SIible . (If a Char Ul OuseQuarwmaster is assigned , . . . . no. . - ityfor lie QMOWto leavethe bridge unless pr~ reliMd.)
The control of the ship is a partially closed information loop . The ' conning officer senses the ship s situation in the world by looking out the window of the bridge . The members of the navigation team also sense the world by looking at it ; in addition , however , they gather information from other sources, and from that other information they synthesize a more comprehensive and accurate representation of the situation of the ship . The navigation team uses its representation to generate advice to the conning officer , who by acting (or not acting ) on that advice affects the actual situation of the ship in the world which is sensed and interpreted . The navigation team relies on the conning officer to the extent that if the conning officer turns the ship or changes its speed in other than the recommended places then the workload of the navigation team is increased . When the quartermasters project the position of the ship into the future , the projections sometimes involve changes in course and or speed. When this is the case, the projected track is carefully planned , precomputed , and plotted . If the ship remains on the precomputed track , many parts of the required computation will have been performed in advance. When the ship deviates from planned track , new computations may be Chapter1 40
required to establish when and where various maneuvers are appropriate ' . For example , on one of the Palau s departures from port an inexperienced conning officer made several turns before the recommended point . This happened because the deck of the ship is so big and so high off the water that from the point of view of the navigation bridge the surface of the water for several hundred yards in front of the ship is hidden from view . When a channel is narrow and some of the turns are tight , channel buoys disappear beneath the deck before the turn is commenced . For an inexperienced conning officer , the temptation to turn before the buoy disappears under the bow is great. Once a buoy disappears beneath the deck, it is difficult to estimate whether or not the ship will hit it . To keep the ship on track , a conning officer must be disciplined and must trust the navigation team. The conning officer has other obligations and cannot always do ' what is easiest for the navigation team. On one occasion the Palau s engineering department detected a rumbling noise in the propeller shaft. In order to diagnose the problem , the engineers requested 50 right rudder , then 50 left rudder , then 100 right rudder followed by 100 left rudder . The ship was slaloming along through 800 turns . This happened while the ship was out of visual and radar range of land , so its position had to be maintained by dead reckoning , a very difficult task under these conditions .
THE COMBATINFORMA nON CENTER The navigation team also coordinatesits activities with the Combat Information Center (CIC), which is located below the flight deck. Duplicate position plots are maintained by the Operations Specialists (OSs) who work in GIG. They use radar bearings and ranges to fix the position of the ship . Under conditions of reduced visibility , CIC is supposed to be the primary source of navigation advice for the conning officer . The quartermaster chief in charge of the Navigation Department on the Palau said the following about this shift in responsibility : ' They ve got a whole team down there [in GIG] and they are pretty good at what they are doing . They are supposed to be like a backup ' on what happens up here. They ve got good radars , and for reduced visibility , they are supposed to be primary . Now the only way that is going to happen is if I drop dead. As long as I am on a ship , and this is the same thing I tell my navigator , as soon as I walk on Welcome Aboard 41
" board , Evel Yfhing that has to do with navigation while I am on ' ' ' board , I m it . I ll hand you papers to sign, I ll back you up in any way you need. You will never get in trouble , navigation is my business " . For OSs, it is a secondary business to them . There are people ' in my business who will let GIG take it . I won t. I never saw this claim put to the test.
AIR BOSS The Navigation Department provides position information to the Air Boss, who is responsible for conuolling the aircraft that operate from the flight deck. The most frequent requests for information from the air boss consist of position or projected position information to be used by aircraft coming to the ship , and directions and distances to land bases for aircraft departing the ship . s. 8IdAnchor Det8I Guiding a large ship into or out of a harbor is a difficult task. A ship is a massive object; its inertia makes it slow to respond to changes in propeller speed or rudder position . Putting the rudder over will have no immediate effect, but once the ship has started turning it will tend to continue turning . Similarly , stopping the engines will not stop the ship . Depending on its speed, a ship may coast without power for many miles . To stop in less distance , the propeller must be turned in the reverse direction , but even this results in only a gradual slowing . Because of this response lag, changes in direction or speed must be anticipated and planned well in advance. Depending on the characteristics and the velocity of the ship , the actions that will bring it to a stop or turn it around may need to be taken tens of seconds or many minutes before the ship arrives at the desired turning or stopping point . ' In order to satisfy the OO D s need for information about the location and movement of the ship when it is near hazards, the Navigation Departments of Navy ships take on a watch configuration called Sea and Anchor Piloting Detail . Piloting waters are defined as follows in the Watch Standing Procedures : Pilotilgwatn - widlilftw mi. of land, ~ or hazardsto navigation, or insideof thefifty fittliilii ClIVI, whicheverIsfurther from land . Resblctedwal88 -Insideof .,. 0II16I" ~ aidto navigationor inside of theten fa UI O In curve, whicheverIsMIler fromIn . Chapter 1 42
1. ~ 0pe I'ati1gwIUIi1 ~ tedWain , U1eSea n AnchorPiloting Detail will be stationed . 2. TheQMOW will . - n Blatall n~ of U1eSea n AnchorPloting Det81 are called at leastthirty minutes prior to 8It8i rlltrictedwatn . Ig - 3. TheSea n AnchorPiloting Detail will cc ;.~ at L TheNavigator b. The~ " I' tothe Navigator c. NavigatlooPlott . d. NavigatlooBearilg Record eli/TIm. - 8. StarboardP8 Ion JS Operator f. PortPelorus OP81tor g. R6.-Ig1!C18dM. - . Y8ringH8hn Im81 h. 0u&t1il1l1 Uterof the Watch i RestiRt IaciM8 * lwriIg H8n1n81kI After Steering j. Fd1 Oln8tlrOP8ator
As long as the visibility is adequate for visual bearings , the primary work of the sea and anchor piloting detail is to fix the position of the ship by visual bearings . The pelorus operators stationed on the port and starboard wings , just outside the doors to the pilot - house, measure the bearings of specified landmarks and report the bearings to the bearing recorder /timer (henceforth referred to as " " the recorder ), who records them in the bearing log. The recorder stands at the after edge of the chart table in the pilothouse . The bearing log is kept on the chart table , adjacent to the chart . The navigation plotter stands at the chart table and plots the recorded bearings as lines of position on the chart , thus fixing the position of the ship . The plotter also projects the future positions of the ship , and together with the recorder he chooses landmarks for the pe- lorus operators to use on future fixes . The restricted -maneuvering helmsman stands at the helm station in the center of the pilothouse and steers the ship in accordance with commands from the conning officer . In sea and anchor detail , the quartermaster of the watch is ' mainly responsible for maintaining the ship s log, in which all engine and helm comm ~ ds and other events of consequence to the navigation of the ship are recorded . The quartermaster of the watch stands at the forward edge of the chart table and keeps the ships log on the chart table . The restricted -maneuvering helmsman is stationed in the after steering compartment , at the head of the rudder ' post in the stem of the ship . In case of a problem with the ship s wheel , the steering function can be taken over more directly by the helmsman in aftersteeringThefathometer operator is stationed in the charthouse , which is separated from the pilothouse by a bulkhead . The fathometer operator reports the depth of the water under the ship for each position fix . The navigator is responsible for the Welcome Aboard 43
work of the navigation team but does not normally participate directly in that work . Aboard the Palau , even the supervision of the navigation team was done by the quartermaster chief , who acted as Assistant to the Navigator . If the crew had been more experienced , the Assistant to the Navigator would not have taken up a functional role in the performance of the task. Because the Palau was understaffed and the available personnel were inexperienced , however , the assistant to the navigator also served as navigation plotter .
N&lliti "w..: ~ ng In the late afternoon of a clear spring day the U.SiS. Palau completed several hours of engineering drills that left it alternately steaming in tight circles and lying dead in the water . The Palau had been at sea for a few days on local maneuvers and was now just south of the entrance to San Diego Harbor . The crew was anxious to go ashore, and going in circles and lying dead in the water when home was in plain sight was very frustrating . It was therefore something of a cause for celebration in the pilothouse when the engineering officer of the watch called the bridge on the intercom " " and said Main engine warmed , ready to answer all bells . The officer of the deck acknowledged the ready state of the propulsion " " plant and advised the engineering officer to stand by for 15, meaning that they should be prepared to respond to an order for 15 knots of speed. Shortly thereafter , the conning officer ordered the engine ahead standard speed. Pilothouse morale rose swiftly . Quartermaster Second Class (QM2 ) John Silver stood at the chart - table in the pilothouse . He was wearing a sound powered telephone set (headphones and a collar -mounted microphone ) that connected him to other members of the navigation team who were not in the pilothouse . When he learned that the ship would be getting underway again soon, he pressed the transmit button on his " ' ' " microphone and said We re baggin ass! On a platform on the starboard side of the ship , just outside the door to the pilothouse and about 50 feet above the surface of the water , Seaman Steve Wheeler had been leaning on the rail , studying the patches of foam that lay motionless next to the hull , and wondering when the engineering drills would end and the ship ' would move again. When he heard Silver s exclamation in his headphones , he looked up and began to scan the city skyline for major landmarks . Wheeler was the starboard pelorus operator , and Chapter1 44
it was his job to sight landmarks and measure their direction from the ship . A novice , he had done this job only once before, and was not sure how to identify all the landmarks , nor was he entirely clear on the procedure he was to perform . Inside the pilothouse , Quartermaster Chief Rick Richards moved to the forward edge of the chart table and looked over the shoulder ' of QM2 James Smith as Smith recorded the conning officer s orders " in the deck log. Ahead standard , left 10 degrees rudder , come to course 3 0 5." Chief Richards turned and leaned over the chart table with QM2 Silver . As happy as they were to be heading for their pier at last, they also knew it was time to begin the high -workload job of bringing the Palau into port . They examined the chart of the ap- proaches to San Diego Harbor . Silver found the symbolic depictions of several important landmarks on the chart and used his fingers to draw imaginary lines from them to the last charted position of the ship . These imaginary lines represented the lines of sight from the ship to the landmarks . He checked the angles at which the lines intersected . Pointing to the chart , he said to " " " " Richards How about these? Yeah, those are fine , the chief replied . ' Silver was the navigation team s bearing recorder . It was his job to control the pelorus operators on the wings of the ship and record the measurements they made. Once Silver had chosen his landmarks " , he assigned them to the pelorus operators : Hey Steve, ' you ll be keeping Hotel del and Dive Tower as we go in , and John, " " " you got Point Lorna. Steve Wheeler answered OK and heard his " " opposite number on the port wing , Seaman John Painter , say Aye . Wheeler looked out across the water , found the conical red roofs of the Hotel del Corona do on the beach, and searched to the south along the strand for the building called the Dive Tower . There it ' was. Wheeler s hands were resting on the alidade that was mounted on a shoulder -high pedestal at his station . He quickly pointed the alidade in the rough direction of the Dive Tower and leaned down , pressing his right eye against the rubber eyepiece to look through the sight . He saw the beach and some low buildings ' back from the water s edge. He swung the sight left and then right until the Dive Tower came into view , then carefully rotated the sight on its pedestal until the vertical hairline in the sight fell right down the middle of the tower . Near the bottom of his field of view Welcome Aboard 45
through the alidade , he could see a portion of the scale of a gyrocompass card. The hairline crossed the scale three small tick marks to the right of a large mark labeled 030. Another large tick mark , labeled 040, was still further to the right . Wheeler counted the little tick marks and noted that the the Dive Tower bore 033 . Once Silver had assigned the landmarks to the pelorus operators , he wrote the name of each of the chosen landmarks at the head of a column in the bearing record log, which was lying on the chart table between him and the chart . Silver kept an eye on his wristwatch . It was a digital model , and when he had come to his duty station several hours ago he had ' synchronized it with the ship s clock on the wall at the back of the pilothouse . Now he had taken the watch off his wrist and placed it on the chart table in front of him , just above the p~ges of his bearing record log. As the ship began to move and turn to its course for home , the plotter , Chief Richards , told Silver to take a round of bearings. It was 13 minutes and 40 seconds after 4 pm . Silver decided to make the official time of the next set of bearing observations - 16 : 14, using the 24 hour notation standard in the military . He " " wrote 1614 in the time column of the bearing record log, and at " " 16 : 13 : 50 he said into his phone set: Stand by to mark . Time 14. Seaman Ron White sat on a high stool at the chart table , looking at the display of the fathometer . On the chart table in front of him " was a depth sounder logbook . When he heard the Stand by to " mark signal in his headset, he read the depth of the water under " the ship from the display and reported on the phone circuit : Fifteen " fathoms . He then logged the time and the depth in his book . Silver recorded the depth in the bearing record log. Out on the starboard wing , Wheeler heard the recorder say " " Stand by to mark , time 14. As he made a small adjustment to bring the hairline to the center of the Dive Tower , he heard the fathometer operator report the depth of the water under the ship as 15 fathoms . The hairline now crossed the scale at 034 . Wheeler pressed the button on the microphone of his phone set and reported " " Dive Tower , 0 3 4. That was a mistake . The bearing was correct ; however , in his excitement Wheeler blurted out his bearing ' immediately after the fathometer operator s report . He was supposed to track the landmark and report its bearing only after the " " recorder gave a mark signal . The port pelorus operator noticed " ' ' ' " the mistake and barked , He didn t say Mark . Chapter1 48
' But by then it was time to mark the bearings. Wheeler s mistake was not a serious timing error ; he was only a few seconds early . The important thing was to make the observations as close the " " mark time as was possible . Stopping to discuss the mistake would have been more disruptive than continuing on. There was no time for lessons or corrections now . The bearing recorder quickly restarted the procedure from its current state by giving the " " mark signal , acknowledging the premature bearing , and urging " the pelorus operators to get on with their reports : Mark it . I got " " " Dive Tower , Steve. Go ahead. Silver then wrote 034 in the column labeled Dive Tower in the bearing record log. The plotter , Chief Richards , was standing next to Silver , waiting for the bearings. He leaned across the chart table and read the bearing of Dive Tower even as Silver was writing it in the log. Silver noticed that Richards was craning his neck to read the bearing " " from the book . Softly he said 0 3 4 to Richards , whose face was close to his . As Richards moved away from the bearing log, he " looked to the plotting tool in his hands and acknowledged : Ub huh ." Chief Richards held in his hands a one-armed protractor called a haey. The hoey has a circular scale of 180 degrees on it , and a straight -edged arm about 18 inches long that pivots in the center of the scale. It is used to construct lines on the chart that correspond to the lines of sight between the ship and the landmarks . Richards aligned the straight edge with the fourth tick mark to the right of the large mark labeled 030 on the scale of the hoey and turned a knob at the pivot point of the arm to lock its position with respect to the scale. He then laid the hoey on the chart and found the symbol on the chart that represented the Dive Tower . He put the point of his pencil on the symbol on the chart . Holding it there , he brought the straight edge up against the pencil point . Keeping the straight edge - against the tip of the pencil and keeping the protractor scale further away from the charted location of the landmark than the anticipated location of the fix , Richards slid the hoey itself around on the chart until the directional frame of the protractor scale was aligned with the directional frame of the chart . The edge of the arm now lay on the chart along a line representing the line of sight from the ship to the landmark . Richards held the hoey firmly in place while he removed his pencil from the symbol for the landmark and drew a line segment along the protractor arm in the vicinity of the expected location of the ship on the chart . By drawing only the sec- WelcomeAboard 47
tions of the lines of position that were in the vicinity of the expected location of the ship , Richards kept the chart neat and avoided the creation of spurious mangles formed by the intersection of lines of position from different fixes . While Chief Richards was plotting the line of position for the Dive Tower , the port wing pelorus operator reported the bearing of Point Lorna. By the time Silver had acknowledged the port pelorus ' " " operator s report ( Three three nine , Point Lorna ), Richards was ready for the next bearing . Because he was standing right next to Silver , he could hear everything that Silver said into his phone - circuit microphone . He could not hear what the pelorus operators or others on the circuit were saying to Silver or to one another ; however , he could hear what Silver said, and he got the bearing to ' Point Lorna by hearing Silver s acknowledgement . -- " While the port pelorus operator was making his report , Point " Lorna, 3 3 9 and while Chief Richards was plotting Dive Tower , Seaman Wheeler swung his sight to the tallest spire of the Hotel del Corona do, aligned the hairline , and read the bearing from the scale. " In his headset he heard the recorder acknowledge Three 3 9, Point " Lorna. But he was ttying not to listen , because he had his own numbers to report as soon as the phone circuit became quiet : " " Hotel del , 0 2 4. Then he listened as the recorder acknowledged " " his report : 0 2 4, Hotel del . The report was heard and echoed without error , so Wheeler said no more . " " About 30 seconds passed between the Stand by to mark signal and the acknowledgement of the third bearing . The pelorus operators relaxed at their stations for a minute or so while the bearings they had reported were processed by other members of the navigation team to determine the position of the ship at the time of the observations . The pelorus operators themselves did not know exactly what had been done with the bearings after they had reported them . Less than 10 seconds after the acknowledgement of the last bearing , Chief Richards had his fix mangle constructed and was ready to label it with the time of the observations . He asked Silver " " ' ' OK, what time was that ? Silver looked in the time column of the " " bearing record log and replied One 4, meaning 14 minutes after the hour . With the fix plotted and labeled , Richards and Silver turned to the tasks of predicting the position of the ship at the time of the next fix (3 minutes hence) and deciding which course to take for
N. ' ;gafD' as CClI~ 1ta1i Dll
Navigation is the process of directing the movements of a craft from one point to another . There are many kinds of navigation . This chapter lays the foundation for the construction of an analysis of the information processing carried out by those who practice a form of navigation referred to in The Western technological culture as surface ship piloting . Piloting (or pilotage ) is navigation involving determination of position relative to known geographic locations . Rather than present what passes in our cultural tradition as a description of how pilotage is done, this chapter attempts to develop a computational account of pilotage . This account of pilotage overlaps portions of the computational bases of many other forms of navigation , including celestial , air , and radio navigation . Aspects of these forms of navigation will be mentioned in passing, but the focus will be on the pilotage of surface vessels in the vicinity of ' ' land . Unless otherwise indicated , the term navigation will hence- forth refer to pilotage . Having taken ship navigation as it is performed by a team. on the bridge of a ship as the unit of cognitive analysis , I will attempt to - apply the principal metaphor of cognitive science cognition as computation - to the operation of this system. In so doing I do not make any special commitment to the nature of the computations that are going on inside individuals except to say that whatever happens there is part of a larger computational system. But I do believe that the computation observed in the activity of the larger system can be described in the way cognition has been traditionally described - that is, as computation realized through the creation , transformation , and propagation of representational states. In order to understand navigation practice as a computational or information -processing activity , we need to consider what might constitute an understanding of an information -processing system. Working on - vision but thinking of a much wider class of information processing systems, David Marr developed a view of what it takes to understand an information -processing system. The discussion here ' is based on Marr s (1982) distinctions between several levels of description of cognitive systems. Chapter2 50
M8r" Levelaof~ --' 4JtIon In his work on vision , Man suggests that there are several levels of description at which any information -processing system must be understood . According to Man , the most important three levels are as follows : The first level is the computational theory of the task that the system performs . This level of description should specify what the system does, and why it does it . It should say what constraints " are satisfied by the operation of the system. Here, the performance of the [system] is characterized as a mapping from one kind of information to another , the abstract properties of this map- ping are defined precisely , and its appropriateness and adequacy " for the task at hand are demonstrated (Man 1982). Such a description is defined by the constraints the system has to satisfy in order to do what it does. The second level of description concerns " the choice of representation for the input and output and the algorithm to be used to transform one into the other ." This level specifies the logical organization of the structures that encode the information and the transformations by which the information is propagated through the system from input to output . The third " level concerns the details of how the algorithm and representation " are realized physically . Marr points out that there are many choices available at each level for any computational system, and that the choices made at one level may constrain what will work at other levels . Marr intended his framework to be applied to cognitive process es that take place inside an individual , but there is no reason, in principle , to confine it to such a narrow conception of cognition . In ' this chapter I will attempt to apply Marr s prescription to the task of navigation . Navigation is an activity that is recognizable across cultures , yet in each cultural tradition it is accomplished within a conceptual system that makes certain representational assumptions . In the next section , I give a computational account of navigation that is independent of the representational assumptions of any established tradition of navigation practice . It is an account that specifies the nature of the navigation problem and the sorts of information that are transformed in the doing of the task, yet spans the differences between even radically different traditions of navigation . Unfortunately , the computational account by itself is quite abstract and difficult to convey in the absence of examples that embody the satisfaction of the constraints that are described . I will Navigation as Computation 51
therefore illustrate aspects of the computational account with a few examples taken from the Western tradition of piloting . This should help the reader to understand the nature of the constraints discussed . However , these examples are inevitably grounded in the representational assumptions of the Western cultural tradition , and frequently have implications for algorithms , and will probably suggest particular implementations . The inclusion of this sort of material seems unavoidable . I will try to keep the examples as sparse as is possible and to make clear distinctions between those aspects that properly belong to the computational account and those that belong to other levels of description . The importance of keeping the computational description free of representational assumptions will become apparent in the two subsequent sections , which briefly contrast the culturally specific representations and algorithms used by our technological Western culture with those used by a nonliterate Micronesian culture to solve the navigation problem . Much of the remainder of the book can be seen as a further elaboration of the representational /algorithmic level of description and a thorough exploration of the implementation of navigation computations by navigation teams on large ships . The implementational details have been largely ignored in the past. This may be due in part to the notion that ininformation - processing systems what is important is the structure of the computation , not the means of implementation . One of the most important insights of computer science is that the same program can run on many different machines - that is, the same computation can be performed many different ways . When we consider a system like ship navigation , however , the situation is complicated by a nesting of computational systems. What is the implemen - tational level for the navigation system as a whole is the com- putationallevel for the people who operate the tools of the system. The material means in which the computation is actually performed are implementational details for the system, but they set the task constraints on the performance of the navigation staff. The distinction between what is computed by the system as a whole and what is computed by the individual navigation practitioners in the system will be developed in later chapters . For the moment , let us take it simply as a justification for attending to a level of detail that is often missing from accounts of organizations as computational systems. Chapter2 52
AComputational AcCOI. !dof N~ 8tion In a computational sense, all systems of navigation answer the " " question Where am 11 in fundamentally the same way . While the representational assumptions of the navigation systems in which this question is answered are enormously variable and wonderful in their ingenuity , all of them answer the question by combining one-dimensional constraints on position . The surface of the sea is, of course, actually a three-dimensional surface on a nearly spherical body , the earth. As long as we are concerned only with positions on this surface, we need only two dimensions to uniquely specify a position . Thus , a minimum of - two one dimensional constraints are required to specify positions for ship navigation . Navigating in three dimensions - a rather recent activity - requires at least three one-dimensional constraints to specify position .
Li8 of POI Iaon - Figure 2.1 depicts the one dimensional constraint that is produced by a known position and a given direction . Such a combination produces a line of position . Thus , if we know that point B lies in a particular direction from known position A , we know that B must lie on a line extended from A in the specified direction . Given that ' constraint alone, however , we still don t know where point B actually is; we know only that it must lie on the line of position defined by point A and the specified direction . If , for example , we are told that a treasure is burled due east of a certain split rock , the ' options are consider ably narrowed but we still don t know where to dig . N Position'
Line of position > - - DirectionDir Actian
Uneofposition Known position .---.-... FIg&n 2.1 Graphicaland CO I1cepb J II depiction ofthe line -ol-Position~ nt. Navigation as Computation &1
crcIII of ~ ~ , Figure 2.2 shows another type of one-dimensional constraint . This one consists of a known position and a specified distance , and it defines a circle of position . If we know that point B lies some specified distance from point A , then we know that B must lie on a circle of position centered on A with a radius of the specified distance . Given this constraint alone, we cannot yet locate B; we know only that it is somewhere on the circle of position specified by the known point and the distance from the point . In practice , a circle of position is often plotted as an arc in the vicinity of the expected location of point B rather than as a complete circle .
Con MIk Ig ~. .~ COI. b' MdI: PosItionFIxiII One-dimensional constraints can be combined in many ways to produce two -dimensional constraints on position . Figure 2.3 shows some of the possibilities . In the Western tradition , the line -of- position constraint is the computational basis of position fixing by visual bearings and by radio direction finding (figure 2.3a). In these procedures , position is determined by finding the intersections of two or more lines of position . A radar fix is constructed from a bearing and a range (figure 2.3b). The circle -of-position constraint is the basis of celestial navigation , although the circles of position
Arcof Position positionArc of position > - - Distan~
Known position fig . - 2.2 Graphicaland conceptual depiction of III arc-of-poeitionconstraint .
LOP AOP AOP Fix Fix Fix LOP> - LOP> - - AOP> - -
a b c FIg In2 .3 ACO I1ceptuaidepiction ofthe combinations ofone -dimensional constraints . Chapter2 54
are so large that they are treated as lines of position in the vicinity of the fix (figure 2.3c). In a celestial sight reduction , each observed celestial body defines a circle of position , and the vessel from which the observations were made must be located at the intersections of the circles of position established with respect to celestial bodies . Systems such as Loran , Decca, and Omega measure time or phase differences between the arrival of signals from multiple stations . Consider position fixing by Loran . If stations A and B emit signals at precisely the same time , where must I be if I receive the signal from station A 3 microseconds before I receive the signal from station B1 The answer is that I must be somewhere on a hyperbolic line of position that is defined by all the intersections of circles of position around A and B for which the circle of position around station A is 3 microseconds closer to station A (at the speed of light ) than the circle of position around station B is to station B. Each pair of stations received provides a time difference that defines ' a hyperbolic line of position . The vessel s position is fixed by finding the intersection of two or more such one-dimensional lines of position . Radar combines a circle of position , expressed as a range (distance ), with a line of position , expressed as a bearing (direction ), to provide a two -dimensional constraint on the relative position of the object detected.
- nil Posl Uon-DIep I8 C8 I71i1itCGI ~ nt " Two other important questions in navigation are Given that we are where we are, how shall we proceed in order to arrive at a particular " " somewhere else? and Given that we are where we are, where shall we be if we proceed in a particular way for a particular period " of time ? Both of these questions concern relationships among positions . To answer the first is to use the specification of two positions to determine the relationship between them . To answer the second is to use the specification of a position and a positional relationship to determine the specification of another position . Both of these constraints are captured by a single constraining relationship that holds among positions and the spatial displacements that lie between them . Figure 2.4 describes this constraint . It simply says that the specification of any two of the items in the relationship fully constrains the specification of the third item . There is no commitment to representation or algorithm in this . Positions Navigation as Computation 55
Position Position
Displacement (direction and distance) fig. . . 2.4 A conceptualdepiction of the positionand displacement constraint .
and displacements may be represented in a wide variety of ways ; however , if they are to be part of a system that does navigation , they will have to be represented in a way that satisfies this generic constraint . Things work out especially nicely if the displacement is given in the form of a direction and a distance . Then the determination of a new position from a given position and a displacement is simply the familiar case of combining the one-dimensional constraint defined by the starting position and a direction and a second one-dimensional constraint defined by the same starting position and a distance . Let me illustrate the satisfaction of this constraint with two procedures from the Western . tradition , course planning and dead reckoning .
COURSEPLANNING The fact that the specification of any two positions uniquely constrains the displacement that lies between them is the basis of course planning . If I know where I am and where I want to be, how can I determine a plan that will get me where I want to go? In some representational systems it is possible to compute a description of how to move from one position to the other from the description of the displacement between two positions . For example , on some types of nautical charts it is easy to measure the direction (course) and the distance between any two locations represented on the chart . Starting at one point and sailing the specified course for the specified distance will deliver the traveler to the other point . In this case, the representational medium , the chart , has been carefully designed so that an easily obtained description of the displacement between positions is also a description of a plan for getting from one point to the other . We tend to take this property for granted , but it is itself an impressive technical accomplishment . Chapter2 58
To appreciate what a nice property it is that displacement on a chart is a plan for travel , one need only consider how often such is not the case. If the positions are represented as street address es to be looked up in a phone book , for example , it may not be easy to get any description of the displacement between them at all . And if one can construct a description of the displacement from the ad- dresses, unless the places are on the same street it is unlikely that the description by itself will be a useful plan of travel .
DEAD RECKONING The fact that the specification of a position and a displacement uniquely constrains another position is the basis of dead reckoning . In dead reckoning the navigator monitors the motion of the vessel to determine its displacement from a previous position . If the distance ' and the direction of the vessel s travel can be determined , the measured displacement can be added to the previous position to determine the current position . Or a planned future displacement can be added to the current position to determine a future position . Thus , if I know where I started and in which direction and how far I have traveled , I caD compute my position . " " According to Bowditch (1977), the term dead reckoning is derived from deduced or ded reckoning , a procedure (predating ' modem charts) in which a ship s position was computed , or deduced , mathematically from a displacement and a known starting position . Even though modem charts permit simple graphical " " solutions to this problem , the term dead reckoning remains . And even though the representation of information and the procedure used in the computation changed with the advent of modem charts , both the old and the new version of dead reckoning are based on the satisfaction of the position -displacement constraint .
DepII-Con- . Matclli1g There is one additional one-dimensional constraint to consider . Nautical charts sometimes have depth contours indicating lines of equal depth of water . If the depth of the water under a ship can be measured , the position of the ship can be constrained to be over a contour of that depth . This is a one-dimensional constraintal - though the line that defines it is usually not a straight line . The utility of this method depends on the shape of the bottom of the sea in the area. If it is a featureless plain , the constraints imposed by Navigationas Computation 57
measured depth are weak . Almost any location on the chart will satisfy the constraint imposed by the measured depth , because the measured depth is almost everywhere the same. If there are many hills and valleys of nearly equal size, then again there may be many contours in many locations that satisfy the constraint . An inclined plane with a moderate slope is a useful bottom shape for simple contour navigation . A measurement of depth in such an area yields a one-dimensional constraint that is typically combined with other one-dimensional constraints , such as circles or lines of position , to generate an estimated position . Another useful bottom shape is encountered in the Central Pacific , where a uniform abyssal plain is dotted with small raised plateaus called guyots . There , one can hop from guyot to guyot , identifying them by their depths . If the depth -measuring apparatus is more sophistica ~ed and can match changing depth contours against patterns of changing depths rather than simply matching single depth measurements against single contours , then additional features on the bottom may provide additional constraints - enough , in fact , to permit a two - dimensional position determination from depth data alone. A similar sort of positional constraint can be achieved on land through the use of an altimeter and a topographic map of terrain that includes altitude contours .
TheDIstance -Rate- TIN Con Ib' I Int Just one more constraint is required to complete the description of - the computational core of navigation . This constraint relates dis lance , rate, and time . Figure 2.5 shows the form of this constraint . As with the constraint that holds among positions and displace - ments , the specification of any two values uniquely constrains the value of the third . The constraint on distance , rate , and time is
Distance Rate
- Time
Fig In 2.5 A conceptualdepiction of the Distance-rate-timeconstraint . Chapter2 58
often used to determine the distance portion of a planned displacement in dead reckoning . This is a commonly used constraint in the Western cultural tradition outside the realm of navigation as well . It is an important part of logistical planning , for organizations and individuals alike . If I walk 4 milt 's per hour , how far can I get - during a 50 minute lunch break? How long will it take me to drive the 118 miles to Los Angeles if I can average 50 miles per hour ? If ' the circumference of the earth s orbit is 584 million miles , how fast is the earth moving along its orbital track ?
Sllnmllyof Consb' ai1t8 The computational account of navigation consists of four principal consuaints . Two of them provide one-dimensional consuaints on position from a given position and a component of a spatial displacement . The third relates positions and the spatial displacements that lie between them , where a displacement is composed of a pair of one-dimensional consuaints (one a distance and the other a direction ). The fourth consuains the relations among distance , rate, and time as descriptions of the motion of an object. Part of the art of the actual practice of navigation lies in integrating information from many kinds of simultaneous consuaints to produce a single solution that satisfies them all .
Re presentation ll iA88 &11 Ip1io118of Western Navigation This section and the next describe sets of structures that have arisen in the Western cultural tradition in terms of which the com- putational consuaints outlined above are represented . The actual mechanics of the techniques for propagating the consuaints across representational structures will be discussed in detail in a later chapter .
Uritl andFrames of Reference In Western navigation , the units of direction are based on a system of angular measurement . This abstract system consists of a circle composed of 360 equal angular units called degrees. By convention , north is 0 degrees, east is 90 degrees, south is 180, and west is 270 degrees. Traditional magnetic compass es had 32 named com- Navi Rationas Computation 5&
pass points . If the compass rose is oriented to b"ue north and south (as defined by the geographic poles), the directions are called true directions . If the compass rose is oriented to magnetic north or south (as defined by the magnetic poles), the directions are called magnetic directions . The magnetic north pole is currently west of Greenland , about 150 from the north geographic pole ; the south magnetic pole is off the coast of Antarctica , toward Australia , about 220 from the south geographic pole . These differences between the locations of the geographic and magnetic poles cause magnetic instruments to show considerable but largely predictable and compensable errors in some locations . For finer resolution , each degree is subdivided into 60 equal minutes of arc, and each minute is further subdivided into 60 equal seconds. Thus , a second of arc is 1/ 1,296,000 of a full circle . It is not widely realized that the coordinates of geographic position (latitude and longitude ) and the basic unit of distance in modem navigation (the nautical mile ) are based on this same system of angular measurement .
GE O G RAP Ifl C PosmON The coordinate system in which locations on the face of the earth are specified is based on a mapping of this circle onto the earth itself . Every location has a latitude and a longitude . The latitude of a place is its angular distance from the equator. Points on the equator itself have latitude 0 . The north and south poles , which are defined by the axis of rotation of the planet , are each a quarter of a circle away from the equator and are therefore at latitude 90 . Locations in the northern hemisphere are said to have north latitude ; those in the southern hemisphere have south latitude . A geometric plane passed through the earth such that it contains the Planet 's axis of rotation will define two meridians where it intersects ' the earth s surface. Longitude is the angular distance of the meridian of a place from an arbitrarily selected meridian that passes through Greenwich , England .
The Greenwich or Prime meridian defines longitude 0 ; its partner , stretching down the Pacific ocean on the other side of the globe, defines longitude 18 . Locations that lie in the 180 to the west of Greenwich are given west longitudes ; those in the 180 to the east are given east longitude . Positions are given in terms of these two one-dimensional constraints . Global positions are speci- fied in terms of this general system. Specific positions are fixed , as Chapter2 .
described in the examples given in the previous section , by their relation to actual locations in the immediately surrounding local space. The nautical chart is a medium in which the specification of positions can be transformed from the local to the global and vice versa.
THE NAUTICAL ~ The nautical mile , the primary unit of distance in maritime navigation , is based on this system of angular measurement . A nautical mile is one minute of arc on the surface of the earth. Thus , there are 360 x 60 = 21,600 nautical miles around the circumference of the earth. The size of this unit has varied historically with variations in the estimation of the size of the earth . Columbus and Magellan assumed a smaller earth that had 45.3 modem nautical miles per degree of latitude . The earth turned out to be about 32 percent larger than they thought . The statute mile (now established as 5,280 feet in the United States) is a descendant of an earlier Roman mile that was also intended to be 1/ 21,600 of the circumference of the earth. As measurement of the earth improved and previous estimates were found to be in error , there were proposals to change the length of the mile itself and proposals to change the number of miles in a degree. For navigation at sea the easy mapping between position descriptions by angular displacement and the size of the major unit of distance is extremely useful . Having one minute of arc equal one nautical mile simplifies many computations at sea. Since this relationship would have been destroyed by changing the number of miles in a degree, the length of the nautical mile was changed. The modem nautical mile - 6,076.11549 feet- is an attempt to preserve that relationship . However , the modem nautical mile is still an approximation . Because the earth is not a sphere, the length of a minute of latitude varies from about 6080.2 feet at the equator to 6108 feet at the poles. One minute of longitude at the equator is about 6087 feet. The current nautical mile is meters exactly . (Bowditch 1977) . The knot , or nautical mile per hour , is the standard unit of velocity in navigation . This knot ties the circumference of the earth to the angular velocity of the earth . Because an hour is 1/ 24 of a day (a complete revolution of the earth), a point on the surface of the earth at the equator moves to the east at a rate of 1/ 24 of the circumference of the earth in nautical miles per hour . That is, 900 knots . Navigationas Computation 81
Charts In the Western tradition of pilotage , virtually all computations involving position are carried out on nautical charts. While there are many other ways to represent the data and carry out the computations of navigation , the chart is the key representational artifact . The most obvious property of maps and charts is that they are spatial analogies. Positions on a map or a chart have correspondence with positions in a depicted large scale space. That is always true . But charts designed for navigation are something more than this . A navigation chart is a carefully crafted computational device . In algebra and analytic geometry , many computations can be performed on graphs; in fact , graphs are essential in motivating the symbolic manipulations that form the real heart of computation in algebra and analytic geometry . One can compute all the points that lie between any two points by drawing a line between them . Or one can identify all the points that lie at a specified distanced from a given point by drawing a circle with radiusd around the given point . Using graphs for computation , however , introduces errors , because plotted lines are less precise than the abstractions they depict (infinitely small points and truly one-dimensional lines ). The infinite set of points lying between any two given points is accurately and economically represented by the equation of the line that contains the two points (and a range on x or y to constrain points to be between the reference points ), and the set of points lying a specified distance from a given point is accurately and economically represented by the quadratic equation for the circle . Of course, the utility of these representations depends on the subsequent computations that they are supposed to support and the sort of computational systems that are available to carry out the computation . It is essential to realize that a nautical chart is more akin to a coordinate space in analytic geometry than to the sort of simple map I may produce to guide a new acquaintance to my office. All maps are spatial analogies in the sense that they preserve some of the spatial relationships of the world they depict , but navigation charts depict spatial relationships in special ways that support certain specialized computations . A navigation chart is an analog computer . Clearly , all the problems that are solved on charts could be represented as equations and solved by symbol -processing techniques . Plotting a position or a course on a nautical chart is just as much a Chapter2 82
computation as solving the set of equations that represent the same constructs as the plotted points and lines . A chart contains an enormous amount of information - every location on it has a spec- ifiable address, and the relationships of all the locations to all the others are implicitly represented . Finally , charts introduce a perspective on the local space and on the position and motion of the vessel that is almost never achieved " ' directly by any person . Standing over a chart , one has a bird s- " eye view that , depending on the scale of the chart , could be duplicated with respect to the real space only from an aircraft or a satellite . Furthermore , the perspective is that of a spectator rather than that of a participant . This is one reason why establishing the correspondences between the features on a chart and the features in the local space is so difficult . In order to reconcile the chart to the territory , one must imagine how the world that is seen from a location on the surface would appear from a point of view from which it is never seen. The chart depiction assumes a very different . perspective than that of the observer on the vessel. The experience of motion for the observer on the vessel is of moving through a surrounding space, while the depiction of motion on a chart is that of an object moving across a space. This other perspective created by the chart is so compelling that a navigator may have difficulty imagining his movements , especially over large spaces, from the ' traveler s perspective . Conversely , people who have had no experience with maps and charts may find them completely balling .
THE COMPUTATIONAL PROPERTIESOF CHART PROJECTIONS Not all charts are equally useful for all sorts of computations . For - example , compare rhumb line sailing with radio -beacon navigation .
Rhumb -line sailing A rhumb line is a line on the surface of the earth that represents ~ constant direction from some location . Rhumb -line sailing refers to a form of navigation in which one sets a course toOa destination and then maintains a constant heading until the destination is reached. When one is steering a ship by any sort of compass, the simplest route is a constant heading . For this task it is very useful to have a chart on which rhumb lines are straight lines . However , if one were to plot the course that would result from steering a constant heading on a globe rather than on a chart , one would produce a line that Navigation as Computation 83
wraps around the globe and spirals up to the pole . This line is called a loxodrome . The Mercator projection overcomes this problem and transforms the spiral into a straight line . Imagine the transformation in two steps. First , the meridians of longitude that actually converge with one another at the poles of the globe are made parallel to one another , so that they are just as far apart at high latitudes as they are at the equator . This introduces a systematic distortion . At the equator there is no distortion , but with increasing latitude the east-west distance shown between the meridians on the chart exceeds by an increasing margin the distance between them on the globe. At the poles , of course, the distortion is infinite - what was zero distance between meridians where they converged at the pole of the globe " would appear as a finite distance on the chart . To compensate for the effects of this distortion on direction , the parallels of latitude are expanded by the same ratio as the meridians of longitude . At the poles this would require infinite expansion , which is why the poles never appear on Mercator projections . This expansion also results in a distortion of the relative areas depicted on the chart . This distortion is more pronounced at higher latitudes . Thus , while Greenland actually has only 1/9 the area of South America , they appear to have roughly the same area on a Mercator chart . Radio -beacon navigation Radio beacon navigation uses radio antennae that are sensitive to direction . Such an antenna can determine the direction from which it receives a radio signal . By tuning the antenna to a station whose location is known and identifying the direction from which the signal comes, one can establish a one-dimensional position constraint . However , a radio signal does not follow a rhumb line ; it takes the shortest route . These shortest routes are called great-circle routes . A great-circle route is defined by the intersection with the surface of the earth of a plane that contains the center of the earth and the two points on the surface between which the route is to be constructed . Great-circle routes can be approximated by stretching a piece of yarn over the surface of a globe. The meridians of longitude define great circles , and so does the equator . All the circles of latitude other than the equator define rhumb -line courses that are not great circles . While the rhumb -line course from Los Angeles to Tokyo is almost exactly due west (and the heading is constant for the entire trip ), the great-circle route leaves Los Angeles heading to Chapter2 84
the northwest and arrives in Tokyo heading to the southwest . To plot a position from radio -beacon bearings, one would like a chart on which great circles are straight lines . Over short distances, great circles approximate straight lines on all projections ; however , over long distances (and radio signals travel long distances) great circles are significantly different from rhumb lines . There is no chart projection on which both rhumb lines and great circles appear as straight lines . In addition to the properties of having rhumb lines and great circles represented as straight lines , it is easy to imagine navigation tasks in which the following would be desirable chart properties : . true shapes of physical features . correct angular relationships among positions . equal area, or the representation of areas in their correct relative proportions constant scale values for measuring distances Whenever the three-dimensional surface of the earth is rendered in two dimensions , some of these properties are sacrificed . For example , the Mercator projection sacrifices true shape of physical features , equal area, and constant scale values for measuring distances in the interest of providing correct angular relationship and rhumb lines as straight lines . These features are most apparent on charts of ' large areas. As the area of the earth s surface represented by the chart decreases, the differences between projections becomes less noticeable . Chart projections make it clear that different representational systems have different computational properties and permit differing implementations of the computations . For example , it is possible to draw a Bleat circle on a Mercator projection ; it is just very difficult to compute where the points should go. On a Lambert conformal chart it is quite easy to draw a great circle , because on this projection a straight line so nearly approximates a great circle that it is more than adequate for navigational purposes . One can see the work that went into constructing a chart as part of every one of the computations that is performed on the chart in its lifetime . This computation is distributed in space and time . Those who make the chart and those who use it are not known to one another (perhaps they are not even contemporaries ), yet they are joint participants in a computational event every time the chart is used. Navigation as Computation 85
S&ln1n8Y - Large scale space is represented as small -scale space on a chart . The primary frame of reference is the system of earth coordinates . Objects that are unmoving with respect to earth coordinates are given fixed locations on the chart . Every location can be assigned an absolute address in a global coordinate system. Direction , position , and distance are all defined in terms of a single universal framework , established by applying a scheme of angular measurement to the earth itself . A universal time standard in combination with the measurement of distance yields a universal unit of rate of movement . These units are universal in the sense that their interpretations do not change with changing location or circumstances of their use. Directions , positions , distances, and rates can all be represented as numbers , and any of the first three can also be modeled in the small -scale space of a chart . Line -of-position constraints are represented as lines on a chart ; circles of position are represented as circles on a chart ; position -displacement constraints are represented as positions and displacements on a chart . Distance , rate, and time are represented as numbers , and computations of the constraints among them are accomplished by digital arithmetic algorithms . All the major computations in this system are based on procedures that involve measurement (which is analog-to- digital conversion ), followed by digital manipulation , followed by digital -to-analog conversion in the plotting of results on a chart .
RepresentationalA8lnpt1o1. of MicrO Ile8 I8 I Nlvigltion The computational account presented above also describes the computations carried out by Micronesian navigators (Hutchins 1983). Micronesian navigators establish their position in terms of the intersections of one-dimensional constraints . Substantial differences between Western and Micronesian navigation become apparent as soon as we consider the representations and the algorithms that the two cultural traditions have developed to satisfy the constraints of the task. A major problem with earlier Western studies of Micronesian navigation was that the representations used in the performance of Western navigation were assumed to be the most general description . Because they failed to see the computa - tionallevel at all Gladwin (1970), Lewis (1972), Sarfert (1911), and Schuck (1882) attempted to interpret the representations used in Chapter2 .
Micronesian navigation in terms of the representations used in Western navigation , rather than interpreting both sets of representations in terms of a single , more general , computational account . This brief discussion of Micronesian navigation is inserted here in the hope that it will make the importance of the distinction between the computational and representational level of description clearer. I also hope to show that even the most commonplace aspects of thinking in Western culture , as natural as they may seem, are historically contingent . In this light , the organization of systems of cultural representations may become visible and, once noticed , may come to seem much less obvious than before. Furthermore , because the representational and implementationallevels constrain each other more closely than do the computational and representational , it is useful to see the relationship between the representational level and its implementation in cultures that are technologically quite different from each other . For more than a thousand years, long-distance non -instrumental navigation has been practiced over large areas of Polynesia and Micronesia , and perhaps in parts of Melanesia . In Polynesia , the traditional techniques atrophied and were ultimately lost in the wake of contact with colonial powers . Only the Micronesians have maintained their traditional skills , and in the past two decades they have been the wellspring of navigation knowledge for a renaissance of traditional voyaging throughout the Pacific Basin (Finney 1979, 1991; Kyselka 1987; Lewis 1976, 1978). Without recourse to mechanical , electrical , or even magnetic devices , the navigators of the Central Caroline Islands of Micronesia routinely embark on ocean voyages that take them several days out of the sight of land . Their technique seems at first glance to be inadequate for the job demanded of it , yet it consistently passes what " " Lewis (1972) has called the stem test of landfall . Of the thousands of voyages made in the memory of living navigators , only a few have ended with the loss of a canoe. Western researchers traveling with these people have found that at any time during the voyage the navigators can accurately indicate the bearings of the port of departure , the destination , and other islands off to the side of the course being steered, even though all of these may be over the horizon and out of sight . These navigators are also able to tack upwind to an unseen island while keeping mental track of its Navigation as Computation 87
changing bearing - a feat that is simply impossible for a Western navigator without instruments . In the neighborhood of the Caroline Islands , less than 0.2 percent of the surface is land . The surface is a vast expanse of water dotted with about two dozen atolls and low islands . Experienced nav- igators in these waters routinely sail their ou trigger canoes up to 150 miles between islands . The knowledge required to make these voyages is not held by all , but is the domain of a small number of experts. The world of the navigator , however , contains more than a set of tiny islands on an undifferentiated expanse of ocean. Deep below , the presence of submerged reefs changes the apparent color of the water . The surface of the sea undulates with swells born in distant weather systems, and the interaction of the swells with islands produces distinctive swell patterns in the vicinity of land . Above the sea surface are the winds and weather patterns which govern the fate of sailors . Seabirds abound , especially in the vicinity of land . Finally , at night , there are the stars. Here in the Central Pacific , away from pollution and artificial liIht , the stars shine brightly and in incredible numbers . All these elements in the nav- ' igator s world are sources of information . The whole system of knowledge used by a Micronesian master navigator is well beyond the scope of this book . Here I will treat only a portion of the nav- ' igators use of celestial cues. The most complete description of this system comes from the work of Thomas Gladwin , who worked with the navigators of Pu- luwat Atoll in what is now the Republic of Micronesia . Gladwin (1970) divides the pragmatics of Puluwat navigation into three parts. First one must set out in a direction such that , knowing the conditions to be expected en route , one will arrive in the vicinity of the island of destination . Second, one must hold the canoe steady on its course and maintain a running estimate of its position . Finally , when nearing the destination one must be able to locate it and head toward it . One of the most widespread notions employed in Pacific non -instrumental navigation is the concept of the star path . From the point of view of the earth , the positions of the stars relative to one another are fixed . As the earth rotates about its axis , the stars appear to move across the sky from east to west. As the earth moves through its orbit around the sun, the stars that can be seen at night Chapter2 .
(that is, from the side of the earth away from the sun) change. But from any fixed location on the earth , any given star always rises from the same point on the eastern horizon and always sets into the same point in the western horizon , regardless of season. Movement to the north or south does change the azimuth of the rising and setting of any star. Within the range of the Caroline Islands navigator , however , the effects of such movements are small (on the order of 30 or less). A star path , also known as a linear constellation " " (Aveni 1981), is a set of stars all of which follow the same path (Gladwin 1970). That is, they all rise in succession from the same point on the eastern horizon , describe the same arc across the sky, and set into the same point on the western horizon . Star paths are typically composed of from six to ten stars fairly evenly spaced across the heavens (Lewis 1972). Thus , when one star in the linear constellation has risen too far above the horizon to serve as an indication of direction , another will soon take its place . In this way , each star path describes two directions on the horizon , one in the east and one in the west , which are visible regardless of season or " " tim ~ of night as long as the skies are clear . A connect the dots drawing of such a linear constellation is simply an arc across the sky, anchored at fixed azimuths in the east and in the west . While the stars themselves make their nightly journeys across the sky, the arcs of the linear constellations remain stationary . Seeing the night sky in terms of linear constellations is a simple representational artifice that converts the moving field of stars into a fixed frame of reference . This seeing is not a passive perceptual process. Rather, it is the projection of external structure (the arrangement of stars in the heavens) and internal structure (the ability to identify the linear constellations ) onto a single spatial image. In this superimposition of internal and external , elements of the external structure are given culturally meaningful relationships to one another . The process is actively constructive . The positions of a few stars may suggest a relationship which , when applied , establish es the identity of yet other stars. Anyone who can identify the traditional Western constellations knows that , in the subjective experience of this seeing , not just the stars but the constellations themselves seem to be " " out there . The little lines holding the stars together seem nearly visible in the sky. These relations are expressed in verbal formulas . " For example , the formula Follow the arc (of the handle of the Big
Navigation as Computation .
Q : SETTING . RISING
POSITIONS POSITIONS
~ ~
\ ~
I
~
~ ~
% '
l4J l4J
$
~
. a : : : ' g : .
~ ~
, . . , ~
~ -
i
~
: j ''4 ;
) ~(,~
~
:
~
\
lE .
At "
AU ) EBAR
ALoE. . QlALAE TAlR
~ :
~
A E AAWl~ TAlRLAE AWl
~ 'S OR\ 0RI0fI
' s
BE L' BELT
~ ~
~ ~ ~ ~ & ~ I ,--SOUTHERN CROSS ; ~ ~. \~ \\ ~ III \\ \ FIg In 2.8 ACaroline Island sidereal compass .
" Dipper ) to Arcturus , and drive a spike into Spica guides the ob- ' server s eye across the sky, consb"ucting part of a constellation . In sky charts for amateur star watchers , the lines are drawn in on the charts- like mental training wheels - to make the constellations easier to imagine when looking at the sky. It is known that star paths have long been used to define the courses between islands in many parts of Oceania (Lewis 1972). The navigators of the Caroline Islands have combined fourteen named star paths with the position of Polaris (the North Star) to form a sidereal compass that defines 32 directions around the circle of the horizon . Figure 2.6 shows a schematic representation of the Caroline Island sidereal compass. As can be seen, most of the recognized star bearings are named for major stars whose paths intersect the horizon at those points . Those which are not so named are the b"ue-north bearing , named for Polaris which from the Caroline Islands is always about 80 above the northern horizon , and three bearings in the south which are defined by orientations of the Southern Cross above the horizon . Of course, the names given to these stars are not the same as the names given to them in the ~ ~ O \ E . ~ , . . Chapter2 70
- Western tradition , nor are all the constellations grouped in the same way . The cardinal direction in the Micronesian system is east , at the rising point of the star Altair . It is interesting that Altair is " " part of a Micronesian constellation called the Big Bird ( hence ' Gladwin s title East is a Big Bird ). The Western tradition has inherited many of its star names from Arabic roots , and Altair is the brightest star in the constellation Aquiia , the eagle . East was the cardinal direction in the Western tradition (consider the two ' ' meanings of the word orient ) before the advent of the magnetic compass . The inclusion of other stars which travel the same path guarantees that as long as the weather is clear the complete compass is available to the navigator no matter what time of year he is sailing . In fact , a practiced navigator can construct the whole compass mentally from a glimpse of only one or two stars near the horizon . ' This ability is crucial to the navigator s performance , because the star bearings that concern him during a voyage may not be those he can readily see. The star compass is an abstraction which can be oriented as a whole by determining the orientation of any part . the the - During day , orientation of the star. compass can be main tained by observing the star bearings from which the major ocean swells come and the star bearings at which the sun and the moon rise and set. Courses between islands are defined in terms of this abstract ' sidereal compass . For every island in a navigator s sailing range , he knows the star point under which he must sail to reach any other island in the vicinity . Thus , the sidereal compass provides the directional reference in terms of which displacements can be specified . The sidereal compass has a second function in navigation : the expression of distance traveled on a voyage . For every course from one island to another , a third island (over the horizon and out of sight of the first two ) is taken as a reference for the expression of the distance traveled . In the language of Puluwat Atoll , this system of expressing distance traveled in terms of the changing bearing of a reference island is called etak (Gladwin 1970 ). Since he knows the star bearings for all the inter -island courses in his area , the navigator knows the star bearing of the reference island from his point of origin and the bearing of the reference island from his destination ' . In the navigator s conception , this reference island starts out under a particular star (at a particular star bearing ) and moves back Navigation as Computation 71
N 9 8 7 6 5 4 3 2 1
WE
Polowat Ruk
FIg In 2.7 An etakdiagram . Thisd ~ . basedon the workof the " ' ~ pherE . Sarfert, reflectsthe conventionalme U1 Odof nwing the relationshipsbetween islands and star pointsfor a typi- caIvoyage .
abeam of the canoe during the voyage through a succession of star bearings until the canoe reaches its destination , at which time the reference island is under the point that defines the course from the destination island to the reference island . The changing star bearing of the reference island during the voyage is illustrated in figure 2.7. The movement of the reference island under the succession of star bearings divides the voyage conceptually into a set of segments called the etaks of the voyage. Each voyage has a known number of etak segments defined by the passage of the reference island under the star bearings. A fundamental conception in Caroline Island navigation is that a canoe on course between islands is stationary and the islands move by the canoe. This is, of course, unlike our notion of the vessel moving between stationary islands . A passage from Gladwin (1970: 182) amplifies this : Picture yourself on a Puluwat canoe at night . The weather is clear , the stars are out, but no land is in sight . The canoe is a familiar little world . Men sit about, talk , perhaps move around a little within their microcosm . On either side of the canoe, water streams past, a line of turbulence and bubbles merging into a wake and disappearing into the darkness. Overhead there are stars, immovable , immutable . They swing in their paths across and out of the sky but invariably come up again in the same places . You may Chapter2 72
travel for days on the canoe, but the stars will not go away or change their positions aside from their nightly trajectories from horizon to horizon . Hours go by, miles of water have flowed past . Yet the canoe is still underneath and the stars are still above. Back along the wake however , the island you left falls farther and farther behind , while the one toward which you are heading is hopefully drawing closer. You can see neither of them, but you know this is happening . You know too that there are islands on either side of you , some near, some far , some ahead, some behind . The ones that are ahead will , in due course, fall behind . Eve Jything passes by the little canoe- eve Jything except the stars by night and the sun in the day.
Here we have a conceptualization in which the known geography is moving past the navigator , his canoe, and the stars in the sky. Of Ito the side of the course being steered is the reference island . It cannot be seen because of its distance over the horizon , yet the navigator imagines it to be moving back slowly under a sequence of star points on the horizon . Observations of navigators during voyages have shown that the navigators can accurately judge the relative bearing of the reference island at any time during the voyage (Lewis 1972). Since the navigator has not actually seen the reference island at any point during the voyage, his ability to indicate where it lies represents an inference that could not be made in the Western system without recourse to tools . ' Gladwin (1970: 184) describes the Micronesian navigator s use of this judgement as follows : ' When the navigator envisions in his mind s eye that the reference island is passing under a particular star he notes that a certain number of segments have been completed and a certain proportion of the voyage has therefore been accomplished . The navigator uses this information to estimate when he will be in the vicinity of his destination , and therefore when he should start looking for signs of land . Since land -based birds venture as far as 20 miles to sea, seeing them arrive at a fishing ground from land or seeing them depart a fishing ground for land can give information at a distance about the direction in which land lies . This information is available only in the early morning and at dusk , when the birds are moving from or to their island . A navigator who arrives at what he believes to be the vicinity of his destination at midday is Navigationas Computation 73
therefore well advised to drop sail and wait for dusk . The danger of failing to make an accurate judgement of when land is near is that one might be close to land when no indications are available and then sail past and be far away from the destination when homing signs are available . Because traditional Micronesian culture is nonliterate , navigators are required to commit a large body of information to memory . Riesenberg (1972) has documented some of the elaborate mnemonic devices used by navigators to organize their knowledge of geography , star courses, and etak segments. An interesting finding ' of Riesenberg s work is that the memorized systems of knowledge make frequent reference to islands that do not exist . Riesenberg (1972: 20) explains :
In a few instances , when unknown geogmphical features were mentioned and when enough courses from identifiable islands to them have been given, an attempt has been made to locate them by projecting the courses on a chart . The intersections of the projected courses genemlly coincide poorly with known bathymetric features . The role of these phantom islands will be taken up in a later section of this chapter .
SomeAnom I Io I8 I"_~ ~ ~ The history of attempts to understand how the Micronesian nav- igators accomplished their way -finding feats reads like a detective story in which we know who did it but not how it was done. Each of several researchers has provided us with both useful clues and a few red herrings . There is little dispute about the nature of course-keeping with the sidereal compass. Western accounts of the star compass go back at least to 1722 (Schuck 1882), and its use seems relatively easy to observe and document . The most detailed description of the star compass of the Caroline Islands was provided by Good enough (1953). Although his diagram reproduced above as figure 2.6 is, as far as we know , a completely accurate depiction of the stars used by the Caroline Island navigators , and although it gives the first complete tabulation of the azimuths (true bearings on the horizon ) and names of the star points , it contains a potentially misleading distortion that was probably incorporated to make the compass concept more accessible to Western readers. Good enough drew the Chapter2 74
compass as a circular compass rose, the way compass es are traditionally represented in our culture . The original records of native depictions of the star compass, however , are all box-shaped. To date there have been two attempts to explain just how the Caroline Island navigators use the concept of etak to keep track of ' ' their progress on a voyage: Sarferts (1911) and Gladwin s (1970). ' Sarferts (1911: 134) description is rich and compact and bears careful consideration:
In an arbitrary voyagebetween two determined islands, the native captains have still a third island in mind, besidesthe starting point and goal of the trip . For the voyagebetween every pair of islands, this is a specific island. Henceforth, I will refer to this island simply " " as emergency island [No tinsel] corresponding to the purpose that it serves as a last place to flee to in case of extenuating circumstances that make it impossible to reach either the starting point or goal of the trip . This island is placed off to the side of the course. In rare situations the natives established two islands as emergency islands , specifically in such a way that one lies to the left and the other to the right of the direction of travel . ' Riesenberg s (1972) discovery that the reference islands for some " " voyages are phantoms , however , makes the emergency island interpretation unlikely . No navigator would attempt to take refuge in a location known to be devoid of land . Another possibility is that knowing the location of the reference island as well as the origin and destination of the voyage allows the navigator to estimate accurately where many other islands in the area are, so that , should he need to take refuge, a choice based on the existing conditions of the wind and the sea might be made among several possible islands . The specification of the placement of the islands is no doubt important ; but if they were places in which to take refuge, would it not be as well to have two " islands " why just . emergency on the same side of the course? Sarfert continues :
In figure [2.7 of this chapter - E.H.] , the island Biseras, a small island of the Onona atoll , serves as emergency island in the already given voyage from Polowat to Ruk [Truk] . If the emergency island is to fulfill its purpose , the captain must be capable of determining at any moment the direction in which the island lies , and therefore the course to it , from an arbitrary point of the voyage. As far as I Navigationas Computation 75
have experience about it , he . . . does this by rather simple means: 1) The direction of the island Biseras from Polowat as well as from Ruk is known . 2) The native captain may undertake a bearing of the area during the trip by means of calculating the already -traveled distance . This is done with the aid of experience , knowledge of the normal duration of the voyage and with the help of an estimate of the speed that the canoe travels through the water. This last means, the so-called dead reckoning , was also in general used by us for the same purpose before the introduction of the log at the end of the sixteenth century . 3) To determine the bearing of the emergency island from the vantage point of the canoe, the observation must necessarily be done such that , as [figure 2.7J clearly demonstrates , it describes the emergency island Biseras, from the canoe as a visible movement on the horizon in the opposite direction of the voyage. This visible movement of the emergency island appears, with the interpretation of the horizon as a straight line , in direct relationship to the already -traversed distance . If the captain estimates, for example , the covered path as being a quarter of the total voyage length , then the emergency island must have completed likewise a quarter of its visible path along the horizon . If the total length of the visible path totals eight (etak) lines , then after one quarter of the trip they would have reached , accordingly , the third line . By means of this simple calculation , the course to the emergency island is confirmed and the captain is capable of seeking it out . (135) ' The major issue raised by Sarfert s proposed calculation technique involves the method used to express the proportion of the total voyage that has been completed . It is easy enough to imagine " - how the navigator might represent the fact that the emergencyis land must have completed a quarter of its visible path along the " " " horizon , although it is doubtful that proportions like a quarter are involved . But how does the captain compute that he has covered some proportion of the total length of the voyage? Further , the expression of the movement of the emergency island in terms of the proportion of the number of etak segments will work only if the etak segments themselves are all nearly the same size. ' ' Gladwin s descriptive model , like Sarfert s, relates the bearing of the etak reference island to the distance traveled . However , Sarfert believed that the navigator computed the apparent bearing of the Chapter2 78
etak island so that he could take refuge there , whereas Gladwin asserted that the navigator used that apparent position as an expression of the proportion of the voyage completed . Gladwin states: ' When the navigator envisions in his mind s eye that the reference island is passing under a particular star he notes that a certain number of segments have been completed and a certain proportion of the voyage has therefore been accomplished . (184) ' This is similar to Sarfert s proportional -derivation model , but the subtle difference raises an interesting question . What is the nature of the computation ? Is it , as Sarfert maintains , that the navigator uses his estimate of the proportion of the voyage completed to establish the bearing of the reference island , or, as Gladwin maintains , that the navigator uses his estimate of the bearing of the reference island to establish the proportion of the voyage that has been accomplished ? Clearly , these concepts are closely related for the navigator . In practice , not every inter -island course is situated such that there is an island to the side of the course with the desired properties of an etak island . Gladwin notes:
If the reference island is too close, it passes under many stars, dividing the journey into a lot of segments. Worse, the segments are of very unequal length . They start out rather long (slow) and then as the canoe passes close by, they become shorter (fast ) as the reference island swings under one star after another , and then at the end they are long again , a confusing effect. A distant reference island has the opposite effect making the segments approximately equal, but so few in number that they do not divide the journey into components ofa useful size. (187)
The effect of having a close reference island is confusing because when a voyage is divided into segments of very different lengths the estimation of the number of segments remaining is a poor measure of the distance remaining in the voyage. Gladwin describes another situation , also noted by Sarfert, in which this same sort of confusion was bound to arise. In a discussion with the master navigator Ikuliman , Gladwin discovered that for the voyage between Puluwat and Pulusuk atolls , a distance of about 30 miles , the navigator used two etak islands - one to the west of the course and nearby , the other to the east and quite distant : Navigation as Computation 77
This case well illustrates one of the difficulties with the practice : when two reference islands are used in this way, the segments are almost certain to be markedly different in length . Ikuliman was not able to offer a good explanation for using two islands , insisting only that this is the way it is taught . When I pressed him further , he observed dryly that Puluwat and Pulusuk are so close together that a navigator does not really need to use ET AK at all in order to establish his position on this seaway, so in this case my question was irrelevant . (188)
Another feature of the system in use that seems to give rise to the same sort of conceptual difficulty is that the first two and last two segments of the voyages are all about the same length , regardless of the positioning of the reference island relative to the courses and regardless of the density of star points in the portion of the horizon through which the reference island is imagined to be moving . Gladwin states: " " Upon leaving an island , one enters upon the ET AK of sighting , a segment which lasts as long as the island remains in view, usually about 10 miles . When the island has at last disappeared , one enters " " the ET AK of birds which extends out as far as the flights of birds which sleep ashore each night . This is about twenty miles from land , making the first two and therefore also the last two, segments each about ten miles long . Having four segments of the voyage absolute in length is logically incongruous (by our criteria ) with the proportional derivation of the remainder of the ET AK divisions . (188)
Again , the problem with this conception is that it interferes with the computation of the distance remaining in the voyage because it destroys the consistency of the etak segments as units of distance . Gladwin explored this inconsistency with his main informant , the - navigator Hipour who later sailed with Lewis to Saipan and back using the system described here (Lewis 1972, 1976, 1978). Gladwin continues :
When I tried to explore with Hipour how he resolved the discrepancy he simply replied that beyond the ET AK of birds he uses the reference island to establish distance . When I asked how he handled the problem of segments ending in different places , under the two methods , he said he did not see this as a problem . As with Chapter2 78
' " " Ikuliman s answer to my problem over the dual referenceis - lands, this ended the discussion. (189) ' " " The major difficulty with Sarfert s model , and all the problems that Gladwin raised with his navigator informants , spring from the observation that etak segments are unsuitable units for the measurement of distance covered on a voyage. One interpretation of this state of affairs is that what appeared to be a logical organizing principle in navigation may be a useful description in the abstract, but that in the exigencies of use it is not strictly adhered to. Gladwin concludes :
Although ET AK has for us much the quality of a systematic organizing principle or even logical construct , the Puluwat navigator does not let logical consistency or inconsistency , insofar as he is aware of them, interfere with practical utility . (189)
There is, of course, another possible interpretation : that the apparent anomalies result from the unwarranted assumption that the etak segments are units of measurement . The notion that consistent units of measurement are necessary for accurate navigation is one of the fundamental representational assumptions of our system of navigation - so much so, in fact, that it is hard for us to conceive of a system of navigation that does not rely on such units and a set of operations for manipulating them . Yet there is no evidence in the record that the etak segments perform that function , nor is there any evidence of any set of mental arithmetic operations that would permit a navigator to manipulate etak segments as though they were units of distance .
A CO Ilcepul Blild Spot The following revealing incident occurred while Lewis was working with the master navigators Hipour of Puluwat and Belong of Pulusuk . According to Lewis :
On one occas;ion I was hying to determine the identity of an island called Ngatik - there were no charts to be consulted of course- that lay somewhere south -west of Ponape. It has not been visited by Central Carolinian canoes for several generations but was an ETAK reference island for the Oroluk - Ponape voyage and as such, its star bearings from both these islands were known to Hipour . On his telling me what they were, I drew a diagram to illustrate that Ngatik Navigation as Computation 78 OROLUK:': - ~f"~.- ..------"I~'.--"L-PONAPE , '30 Riles / ' FIg In 2.8 Lewiss methodof detenniningthe positionof the islandNgatik .
must necessarily lie where these ET AK bearings intersected . [See figure 2.8.J Hipour could not grasp this idea at all . His concept is the wholly dynamic one of moving islands . (1972: 142) This passage raises several important questions : Why did Lewis use the technique of drawing the intersecting bearings in order to determine the location of the island called Ngatik ? Why did Lewis assume that posing the question the way he did would make sense to Hipour ? Why did Hipour not grasp the idea of the intersecting bearings? Let us consider the questions about Lewis first . The technique Lewis used is clearly an effective one for the solution of this particular problem . It establish es a two -dimensional constraint on the location of Ngatik by combining two one-dimensional constraints . It also contains some very powerful assumptions about the relation of the problem solver to the space in which the problem is being solved . First , it requires a global representation of the locations of the various pieces of land relative to each other . In addition , it requires a point of view relative to that space which we might call the " ' " bird s-eye view . The problem solver does not (and cannot without an aircraft ) actually assume this relation to the world in which the problem is posed. We can guess that Lewis did this because it is for him a natural framework in which to pose questions and solve problems having to do with the relative locations of objects in a two -dimensional space. Western navigators make incessant use of this point of view . When a Western navigator takes a bearing of landmark , he has a real point of view on a real space. However , as soon as he leans over his chart , he is no longer conceptually on the boat; he is over the sea surface, looking down on the position of his craft in a representation of the real local space. Novice navigators - - z - , Chapter2 .
sometimes find this change of viewpoint disorienting , especially if the orientation of the chart does not happen to correspond to the orientation of objects in the world . ' Belong was also puzzled by Lewis s assertion , and in reaching an understanding of it he provides us with an important insight into the operation of the Micronesian conceptual system: He eventually succeeded in achieving the mental tour de force of visualizing himself sailing simultaneously from Oroluk to Ponape and from Ponape to Oroluk and picturing the ET AK bearings to Ngatik at the start of both voyages. In this way he managed to ' comprehend the diagram and confirmed that it showed the island s position correctly . (143) ' The nature of Beiong s understanding indicates that for the Caroline Island navigator the star bearing of an island is not simply the orientation of a line in space but the direction of a star point from the position of the navigator . In order to see that the star bearings would indeed intersect each other at the island , he had to imagine himself to be at both ends of the voyage at once. This allowed him to visualize the star bearing from Oroluk to Ngatik radiating from a navigator at Oroluk and the star bearing from Ponape to Ngatik radiating from a navigator at Ponape. What Hipour probably imagined when Lewis asserted that the island lies where the bearings cross must have been something like the situation depicted in figure 2.9. Contrast this with what Lewis imagined ' he was asserting (figure 2.8). Hipour s consternation is now perhaps more understandable . The star bearings of the etak island radiate out from the navigator himself . From this perspective they meet only at him . In his conception of this voyage, the etak island begins under one of these bearings and ends under the other . That two relative bearings might meet anywhere other than at the navigator is inconceivable . Because the Caroline Island navigator takes a real point of view on the real local space to determine the star bearings, it does not seem likely that the mapping of etak segments onto an abstract representation of the expanse of water between the islands is faithful ' to his conception . Gladwin s (1970) statement about the nav- ' " igator s noting that a certain number of segments have been " completed and the diagrams that Lewis , Gladwin , and Sarfert use to represent the changing relative bearing of the etak referenceis - land all contain two implicit assumptions : that the navigator uses Navigation as Computation 81
ALDEBARAN settin Q
ALTAIR ALTAIR (BIGSIR ) ...... (BIGBRD ) . tting ....---~---"""-'to._.~-~;;;:"I-~- .~"I .-..,."...------~--~-~ -~ ~-~/0""-'~ "-Gto-.':.:.i~-.~-. +-~ - ~-~--...... ~rising CORVIrisinQ settin G ' some sort of bird s-eye view of the space he is in , and that he conceives of a voyage in terms of changes in the position of his canoe in a space upon which he has an unchanging point of view . These ' assumptions are true of the Western navigator s conception of a voyage, but they appear not to be true of the Caroline Island nav- ' igator s conception . These assumptions are at odds with the verbal data (i.e., descriptions of the islands moving relative to the navigator ) and with the behavioral data (i .e., consternation in the face of what ought to be a trivial inference ). It is tempting to criticize the Caroline Island navigators for maintaining an egocentric perspective on the voyage when the global perspective of the chart seems so much more powerful . Before concluding that the Western view is superior , consider the following thought experiment : Go at dawn to a high place and point directly at the center of the rising sun. That defines a line in space. Return to the same high place at noon and point again to the center of the sun. That defines another line in space. I assert that the sun is located in space where those two lines cross. Does that seem wrong ? Do you feel that the two lines meet where you stand and nowhere else? In spite of the fact that the lines seem to be orthogonal to each other , they do cross at the sun. This is not in - tuitively obvious to us, because our usual way of conceiving of the Chapter2 . Night Day Noon - ~- ~ Dawn Sun Earth Figure2.10 A heliocentricdepiction (not to scale) of pointingat the sooat dawnand then again at noon. Thesun Is indeedlocated where the linescross . ' sun s location is not to conceive of its location at all . Rather, we think of its orientation relative to a frame defined by the horizons and the zenith on earth. The rotation of the earth is not experienced as the movement of the surface of the earth around its center , but as the movement of the celestial bodies around the earth . From a point of view outside the solar system, however , the intersection of the lines is obvious , and it is immediately apparent that the sun is in fact located where the lines cross (figure 2.10). ' Our everyday models of the sun s movement are exactly analogous ' to the Caroline Island navigator s conception of the location of the reference island . The choice of representations limits the sorts of inferences that make sense. Because we Westerners have all been exposed to the ideas of Copernicus , we can sit down and convince ourselves that what we experience is an artifact of our being on the " " face of a spinning planet . That is, after all , the correct way to think of it , but it is not necessarily the most useful way . Modern celestial navigation is deliberately preCopernican precisely because a geocentric conception of the apparent movements of bodies on a rigid celestial sphere makes the requisite inferences about the positions of celestial bodies much easier to compute than they would be in a heliocenuic representation . From a perspective outside the galaxy , of course, the heliocentric conception itself is seen to be a fiction which gives an improved account of the relative movements of bodies within the solar system but which is incapable of accounting for the motion of the solar system relative to the " " other stars in the universe . Such a veridical cosmology is irrelevant ' to any present-day navigator s concerns . These observations place suong consuaints on candidate models of how the Caroline Island navigators use the etak system. Viable models must not rely on arbiuary units of distance , nor should they Navigation as Computation 83 ' involve a bird s-eye view of the navigator and his craft situated in some represented space. An Altema Uve Model What does the Caroline Island navigator gain by using the conception of the moving reference island ? Western navigators find the use of a chart or some other model indispensable for expressing and keeping track of how much of the journey has been completed and how much remains . While the Caroline Island navigators are fully capable of imagining and even drawing charts of their island group , these conceptions are not compatible with the moving - island and star-bearing conceptions they use while navigating . ' Lewis s diagram was nonsense to Hipour because Hipour never ' - takes a bird s eye point of view when he is thinking about star bearings. In addition , even though the necessary technology is available to them , we know that the navigators carry nothing like a chart with them on their voyages. ' Consider the Caroline navigator s conception in its context of use. At the outset of any voyage, the navigator imagines that the reference island is over the horizon ahead of him and to one side. It is, for him , under the point on the horizon marked by the rising or setting of a particular line of stars. During the course of the voyage, the reference island will move back along its track , remaining out of sight of the navigator . As it does so, it will assume positions under a succession of star bearings until it lies under the star bearing that marks the course from the destination to the referenceis - land . If the helmsman has kept a straight course, then the canoe will be at the destination when this happens . An important aspect of this imagined sweep of the reference island back along its track , out of sight of the navigator , has been ignored by recent writers on Caroline navigation but was noticed by Sarfert in 1911. Sarfert was struck by the fact that the navigators conceive of the horizon as a straight line lying parallel to the course of the canoe. For a Western navigator , who normally conceives of the horizon as a circle around him , this is a puzzling observation . Why should these nav- igators make such a counterfactual assumption ? Sarfert realized the importance of the fact that the Caroline navigator conceives of the horizon as a straight line and imagines the apparent movement of the reference island beyond it . With this Chapter2 84 Eta~ at end Canoe Canoe Canoe at start at end a b . Movement Movement of canoe figure2. 11 (I) Thestandlrd Weswn repr . . . tationof fie mov6 ITlilit of fie canoe8Id ciwVng star bearingsto fie etakisland . (b) TheMIc .-;;;-.aIian i81ir ~ Iiatk;.. of fie lime ~ ~ 81 fie movemerltof fie eta Ic island. . . fie starbearings . (c) Ilh8tratb. UIatfie I~ ned movementof fie eta IcIsland Is a modelof fie movernentof fie canoealong fie COtne bearing through a set of intermediate bearings to the final bearing is exactly proportional to the progress of the canoe from the island of departure across the sea to the goal island (figure 2.11). Of course, ' the navigator does not think of it from the bird s-eye perspective provided by the figure . Rather, the imagined movement of the etak reference island just under the horizon is a complete model of the voyage which is visualizable (but not visible ) from the natural point of view of the navigator in the canoe (figure 2.12). It is a repre- Navigation as Computation . C<=U~j .-QtU-) .Q~i~-. <-.';Q CDE) (E(tU!ij) <-.S.-. ..m'- Q=)t .0. Qc. .. u~0II;) ~ 2.12 Thehorizon with star points as seen from the canoe . Whenthe navigator looks at the horizon , he imaginesthe locationsof the starbearings . In I Ils diagram, the constellationOrion is shownrising . Thisserves as ananchor for theconstruction of theentire star compass , i I- cludingpoints defined by starsthat are not presently visible . Theshaded ~ on belowthe horizonrepresents the water between the canoe and the horizon . ' sentation of the spatial extent of the voyage, and of one s progress along it , that does not require either the construction of a map or a change of viewpoint . The straight -line -horizon conception is essential to the transformation of angular displacement into linear displacement . The image of the etak reference island moving along just below the horizon can be quite naturally tied to the passage of time . Part of the knowledge that a navigator has about every voyage is the amount of time he can expect the trip to take under various conditions . Suppose that the navigator knows for a particular voyage that , under favorable conditions , he will arrive at his goal after one day of sailing . If he leaves his island of departure at noon (acom - mon departure time ), he can estimate that he will arrive at his destination at about noon on the following day. In terms of the movement of the reference island , this means that the island will move from a position under the initial bearing to a position under the final bearing in one day (figure 2.13). Still assuming a normal rate of travel , he can associate other times during the voyage with other bearings of the reference island (figure 2.14). In so doing , he not only has a visual image that represents the extent of the voyage in space; he also has one that re~resents the voyage and its subparts in time . If the sailing conditions are as expected , the task of determining where the reference island is positioned over the horizon at any point in time is trivial . All the navigator need do is determine Chapter2 . Etakisland here Etakisland here at endof voyage at startof voyage + c + iQC/) as~ ... .cQ-:) <-'.Cg 's< Figure2.13 Thesuperimposition ofstarting and ending-- Ibearing.. son: thestar points . Thestar bearing of theetak island atthe start of the voyage isunder rthestal , pointdefined- byAntares . Atthe end ofthe voyage the star bearing oftheeta~ - -~_ A:island~-~-~~- ~is~under- ~~-_.' the Pleiades. Theetakisimagined to movealong beyond the horizon from the star point defined byAntares tothe star point defined bythe Pleiades . Etakisland here Etakisland here atend of at start of voyage ~Q) voyage + ":5 Q) ue 0- ~ + Sepe O!( . ~ 5 S9 0- snAJO tomorrow Figure2.14 Temporailandrnarkssuperimposed onstar points and the image of the etak island . Theexpected durationofthe voyage is mappeduniformly onto the space defined by the starting andending star bearings ofthe etak island . Jeqepl V ! eld U Jel V ~ Navigation as Computation 87 Etakisland here Etakislandhere at start of voyage atend1,ofvoyage ~~ cuQ ) - C 0- =~ C/) co C/) 'cu- 0- "3 Q) cu ~ co ..c "0cuE 0 .Q!~.!. :- 2g ~ cuE ~ .:m~Q ) 0'CS u V) Noon Sunrise Midnight Sunset Noon tomorrow today figure2. 15 Justbefore midnight the navigator points to theetak island . Allhe needs to dois pointto the locationof thecurrent time on the time scale that is superimposedonthe spatial landmarks providedby the star points . the time of day and refer to the image of the reference island moving along under the horizon . By pointing to the position on the horizon that represents the present time of day, the navigator has pointed directly at the reference island (figure 2.15). The assumption that etak segments are units of distance led Gladwin to three related apparent inconsistencies : the supposedly confusing effect of having etak segments be of different lengths , the conflicting boundaries of etak segments defined by using more than one etak island at once, and the conflicting boundaries of etak segments at the beginning and end of a voyage (caused by using the etak of birds and the etak of sighting in addition to the star-bearing - " defined etak segments). Gladwin found these conceptions completely " inconsistent with the theory as described above (189). In my model , there is no need to assume that the etak segments are units of distance . We dispense with the notion that the numbers of etak segments enter into a numerical computation of the proportion of the voyage completed or remaining . The inequality of their lengths is not an awkward conceptual problem ; it simply means that on a typical voyage the navigator will have more conceptual landmarks defined by star bearings in the middle of the voyage than at the ends. In fact, if we listen to the navigators , we sa J ~ UV . . . Chapter2 . find that they are not talking about the spatial duration (length ) of the etak segments, but of their temporal duration . As Gladwin " ' ' (1970: 187) notes, They start out being rather long ( slow ) and ' ' then as the canoe passes close by , they become shorter ( fast ) as the reference island swings under one star after another , and then at " the end they are long again, a confusing effect. The concern of the navigator is not how far he travels in a particular etak segment, but how long he will travel before asserting that the reference island has moved back under the next star bearing . When the concept of the etak segment is freed from the notion of a unit of distance , the apparent problem of using more than one etak island at once, and the apparent problem of overlapping the - - star bearing determined etak segments with those determined by the range of birds and the range of sighting disappear . Using one etak island to each side of a voyage gives the navigator more con- ceptuallandmarks on his voyage. There is no reason for it to be a problem to the navigator . If two reference islands were on the same side of the voyage, however , the navigator would have two complete - but non co extensive sets of time -bearing correspondences superimposed on a single horizon , and that probably would be a source of confusion . But Sarfert (1911: 134) was quite clear on this issue; he said that when two etak islands are used, they are chosen " specifically in such a way that one lies to the left and the other to " the right of the direction of travel . The confusion that Gladwin imagined with one reference island to each side does not arise, since the etak segments are mapped not onto the course line but onto the imagery on the horizon in front of the reference islands (figure 2.16). The strategy of including the etak of sighting and the etak of birds is entirely consistent with the notion of the star-bearing - defined etak division as a conceptual landmark . The star-bearing - defined etak segments are conceptual landmarks derived in a particular way , and the etak of sighting is a conceptual landmark determined in another way . Once established , they function for the navigator in the same way . They do not enter into a numerical computation ; rather , they give the navigator a more direct representation of where he is (or, actually , where land is). In addition , since the star-bearing etak segments are slow in passing near the beginning and near the end of the voyage, it may be helpful to the navigator to have the other conceptual landmarks at those points . Navigation as Computation. Bearingof#2etak of #1 etak atstartof Bearing voyage at start of voyage Right UOZIJOH ~ Bearingof#2etak Bearing of #1 etak atendofvoyage at end of voyage f9n 2.18 Theelect of &8ingtwo etakI8 Iand8, oneon eachside of a. COtnI. Thestar be Iri1gs of a. twoetak islands do not lit Gt1il~ will eachoiler , ~ -'-~ theyare ~ ontoseparate imagescon . mJCt8don lie horizonson oppositesides of lie course. What of the phantom etak islands that correspond to no known bathymetric features? If the conception of etak presented here is correct , there is no need for there to ever be an island present at the etak point . One need only decide , for any particular voyage, that one is going to model the progress of the voyage as the movement of an unseen point that starts out under a star bearing ahead of and to the side of the course and ends up under a star bearing behind and to the same side of the course. Such a phantom consb"uct does all the conceptual work required of the etak. Neisser has remarked that the error of assuming that etak islands must be safety islands to " which one sails in case of danger is an overly concrete interpretation " of the navigators abstract idea (Neisser 1976, cited in Frake 1985). This conception and this technique make computing the location of land trivial when conditions are favorable . Suppose, however , that a voyage must be made under conditions which differ from those expected at the outset of the voyage. How could the navigator update his image of the movement of the reference island to reflect Horizon Chapter2 . what is happening to his rate of travel ? The key to this problem lies in the judgement of speed and in the way that this judgement is expressed. Any experienced Western yachtsman can make fairly ' accurate judgements of his boat s speed through the water without the aid of instruments . By attending to the feel of the boat as it moves through the water , the accelerations developed as it moves over waves, the feel of the apparent wind , the appearance and sound of the wake (it sizzles at speeds in excess of about 5 mots ), the response of the helm , and many other sensations, the small - boat sailor can make judgements that he normally express es as a number of units - usually mots . The knot is a good choice for the yachtsman ; as one nautical mile per hour , it is convenient for subsequent numerical calculations . One might have expressed the speed as furlongs per fortnight , or on a scale of how thrilling it is, but neither of these fits especially well with useful subsequentcal - culations . The same must be true for the Caroline Island navigators . There is no doubt that they can make accurate judgements of speed; however , expressing those judgements in terms of mots would not be advantageous at all for them , because that unit is not compatible with any interesting computations on a visual image of the moving reference island . Clearly what is wanted is an expression of speed that bears a compatible relationship to the imagery . Consider the following hypothetical scheme. At some point in the voyage (and it could be any point , including the very beginning ) the speed of the canoe changes. The navigator reconstructs his image of the movement of the reference island with the time landmarks placed in accordance with the previous speed. If the change occurs at the very beginning of the voyage, the usual or default speed will be taken as the previous speed. Let the segment of the horizon from the present position of the reference island to any convenient future time landmark " " represent the previous speed (see the segment labeled old rate in figure 2.17). This represents the expected movement of the reference island at the previous speed during the period between the present time and the temporal landmark chosen. The problem is to determine the movement of the reference island during the same time period at the new speed. If the new speed is greater than the old speed, then the reference island will move further along the horizon in the same period ; if the speed is less, the movement will be less. Using the old rate as a scale, imagine another segment Navigation as Computation 91 Etakislandhere Etak island here atend ofvoyage at start of voyage ael + G> - QU) 'tVc- ~"5 "C tV 0- C/) co .tV- .Q.)c L. C :J "'5 .Coo!? "C ~.G5> .0 ~ orco. '< ~m 0~ u0 V) Noon Sunrise tomorrow (newrate ) (newrate ) figure2. 17 Reconstructingtheetak imagery toreflect achange ofspeed . " (" new rate in figure 2.17), starting at the present position of the reference island and extending in the direction of the apparent movement of the reference island . This segment represents a judgement of the magnitude of the new speed relative to the old speed. Now simply move the time landmark from the end of the old -rate segment to the end of the new-rate segment. The new -rate segment now defines the new time scale for the new speed. The other time landmarks for subsequent portions of the voyage can be moved accordingly , as in the figure , and a complete new set of expectations for the times at which the etak reference island will assume future positions is achieved . This procedure can, of course, be applied anytime there is a noticeable change in the rate of travel of the canoe through the water . Thus the navigator can always keep an updated set of time -bearing correspondences for the etak reference island which allows him to gauge how much of his voyage has been completed and how much remains . The notion of the changing bearing of the reference island can be accommodated by our usual way of thinking , in which the canoe is in motion while the islands remain fixed . Why , then , would nb ! ' V ' ewwe9 se J ~ UV ...... Chapter2 12 Micronesian navigators insist on what they know to be a fiction and imagine that the canoe is stationary , with the islands in motion about it ? All navigation computations make use of frames of reference . The most prominent aspect of the Micronesian conception is the apparent motion of the etak island against the fixed backdrop of the star points defined by the sidereal compass. Here there are three elements to be related to one another : the vessel, the islands , and the directional frame. In order to preserve the observed relationships of motion parallax , one can have the vessel and the direction frame move while the islands stay stationary (the Western solution ) or one can have the vessel and the directional frame stationary while the islands move (the Micronesian solution ). In the Western case, the directional frame is a compass, or a gyrocompass , and it is carried with the ship . In the Micronesian case, the directional frame is defined by the star points of the sidereal compass, and the star points are fixed . Each of these schemes makes some things easy to compute and others difficult . The islands move for the Micronesian navigator , because it is computationally less expensive to update their positions with respect to the frame defined by the navigator and the star points than it is to update the positions of both the navigator and the star points with respect to the positions of the islands ( Hutchins and Hinton 1984). S81m8Y The position -displacement consuaint is represented locally in the Micronesian system in every inter -island course. Sailing a constant heading from a known location implicitly represents a line of position . A second line of position is established by the imagined bearing to the etak island . The position of the canoe is established as simultaneously satisfying these two one-dimensional con- suaints , although the two representations are not superimposed ' directly on each other , as they are on the Western navigator s chart . The line of position representing the uack of the canoe is implicit in the steered course of the canoe. The concepts of the etak of birds and the etak of sighting provide a circle of position consuaint . Depth contours are also used, and the Micronesian navigators practice a form of guyot hopping on some voyages by sailing from Navigation as Computation. se amount to se amount . Even though they do not encounter land , they are able to determine their position by the discoloration in the water caused by the presence of the submerged se amount . The - - distance rate time constraint is explicitly represented in the superimposition of temporal landmarks on the spatial landmarks defined by the star bearings of the etak island . In this system there are no universal units of direction , position , distance , or rate, no analog- - to digital conversio ~s, and no digital computations . Instead , there " " are many special -purpose units and an elegant way of seeing the world in which internal structure is superimposed on external structure to compose a computational image device . By constructing this image, the Micronesiannavigatorperforms navigation " ' " computations in his mind s eye . Pll-Mo8n W8It8mNlvigl Uon The practice of modem navigation is of more recent origin than many of us probably imagine . Before the introduction of the magnetic compass (around 1100 AiD .), navigation in European waters looked a good deal like a rather unsophisticated version of Micro - nesian navigation . We do not know the extent to which the sim - ilarities between the two systems are due to independent invention or how much they share from a common origin . Some scholars have attempted to find a common Arab origin for some of the features (Lewis 1976 ), but the evidence of such a connection is scanty at best . Whatever the reasons for their existence , consider the following parallels . Before the discovery of the magnetic compass needle , the sun and the stars were the guides for Western navigation . In the Odyssey , Homer has Odysseus come home from the west by keeping the bear (the Big Dipper ) on his left and sailing toward the rising of the Pleiades and Arcturus . The Pleiades and Arcturus have similar declensions (they rise out of the same point in the eastern horizon ) and are 11 hours different in right ascension (they are on opposite sides of the night sky ), so one or the other would be in the sky on any night regardless of season (Taylor 1971 ). This is clearly a linear constellation construct , although having only two stars in the constellation is of limited utility (since the navigator will not always have one of the stars near enough the horizon to be useful for course setting ). Chapter2 M In ancient Greece, very short distances were given in stadia (a stade is about a tenth of a mile ), but longer distances in early voyages ' were given in terms of a day s sail . This was the distance a " normal ship would accomplish during a twenty -four -hour run " with a fresh following wind (Taylor 1971: 51). The units in which the distances between islands are given in the Micronesian system are based on exactly the same concept , the only difference being ' that Micronesians are interested in a day s sail of a canoe (Riesen- berg 1972). This still requires the navigator to recognize the conditions " ' " under which a day s sail will be accomplished in a day. Making this judgement is probably the sort of skill that no practitioner " can describe in detail - But ever since sailing began, masters ' ' and pilots have always prided themselves on knowing the feel " of their ship and how much way she was making (Taylor 1971: 52). " The kenning , a unit of distance used by early mariners , equivalent to the distance at which the shore could first be seen from the " offing when making landfall (Cotter 1983b: 260), appears to be a European version of the etak of sighting - although , since the decks of European ships are generally higher than the decks of Micro - nesian canoes, it is a greater distance . This is a salient concept for mariners of all kinds . In the Western system it became the basis of a unit of distance . Once determined , it was used as a unit of distance " in sailing directions that give the kennings between headlands " and ports (Cotter 1983b: 255). The sighting of birds has been important in the Western tradition since biblical days. Fuson (1987) reports the following entries in the log of Christopher Columbus on his first voyage to the New World : Later in the day 1 saw another tern that came from the WNW and flew to the SEeThis is a sure sign that land lies to the WNW because these birds sleep ashore and go to sea in the morning in search of food , and they do not fly sixty miles . (65) 1 know that most of the islands discovered by the Portuguese have beenfound becauseof birds. (71) The first quote shows that Columbus was not only using the behavior of birds to find land, he was also making the same sort of inferences as are made by the Micronesian navigators. The second ' quotation gives an indication of Columbuss estimation of the importance of this technique. Since in the century before his voyage Navigation as Computation . no European nation had discovered more islands than Portugal , this is a strong endorsement of the technique . When Europeans first ventured into the open ocean, they could roughly determine latitude by measuring the altitude of the North Star, or of the sun as it passed the local meridian . Yet they had no way to determine their longitude with any accuracy . To find an island known to be at a particular latitude and longitude , a European navigator would attempt to arrive at the target latitude well upwind of the target longitude ; he could then simply sail downwind , maintaining the specified latitude until the island was " " sighted . This technique of latitude sailing was probably practiced by traditional Pacific navigators too , although because of the nature of the traditional practices the evidence is simply lacking . It is interesting to note , however , that a young Hawaii an navigator , Nainoa Thompson , who apprenticed himself to an experienced Caroline Island navigator , has invented or discovered a technique for determining the latitudes of specific islands at sea, and has used - this technique to support the latitude sailing strategy in long - distance voyages between Hawaii and Tahiti without the aid of instruments . The technique relies on the observation of pairs of stars rising out of or setting into the horizon . At a particular latitude , if one can find two stars that rise out of the eastern horizon at the same instant , then the more northerly of the two will rise before the other when the observer is north of that latitude , and the more southerly of the two will rise before the other when the observer is south of that latitude . By identifying a few pairs of stars for each target island , it is possible to use the latitude -sailing strategy with great accuracy . TheDlvergel1 C8of Tracltions The similarities between early European navigation and Micro - nesian navigation are based on regularities in the world that are just too useful to miss. The differences between the two traditions are many and appear to have increased in number over time . The divergence of the traditions can be traced through three closely related trends in the development of Western navigation : the increasing crystallization of knowledge and practice in the physical structure of artifacts , in addition to in mental structure ; the development of measurement as analog-to-digital conversion , and the concomitant relianceontechnologies of arithmetic computation ; Chapter2 . and the emergence of the chart as the fundamental model of the world and the plotted course as the principal computational metaphor for the voyage. - nil CrystallzaaonofK , . . ." 8Id Practiceit lie Pf Ir IlcalSbucue of Artifacts The Micronesian navigator holds all the knowledge required for the voyage in his head. Diagrams are sometimes constructed in the sand for pedagogicalpurposes , but these (of course) are only temporary and are not taken on voyages. In the Western tradition , physical artifacts becamerepositories of knowledge, and they were constructed in durable media so that a single artifact might come to represent more than any individual could know. Furthermore, through the combination and superimposition of task-relevant structure, artifacts came to embody kinds of knowledge that would be exceedingly difficult to represent mentally (Latour 1986). Many of the instruments of Western navigation are basedon the principle of building computational constraints of the task into the physical structure of the artifact. I will illustrate this pervasive strategywith just a few examples. THE ASTROLABE The asuolabe (figure 2.18), a portable mechanical model of the movements of the heavens, was invented in Greece around 200 B.C. Preserved during the Dark Ages by the Byzantines , it was not much modified by the Arabs , via whom it returned to the West around 1000 AiD . An astrolabe is a memory for the structure of the heavens. As we saw in the discussion of Micronesian navigation , it is possible for an individual navigator to learn an internal image of the heavens so rich that he can recognize arrangements of stars, and even imagine the locations of stars that are obscured by cloud or the horizon . However , it is not possible with such mental representations to control all those spatial relationships with the sort of precision that is possible in a durable external representation . In an external representation , structure can be built up gradually - a distribution of cognitive effort over time - so that the final product may be something that no individual could represent all at once internally . Furthermore , the astrolabe encodes a kind of knowledge that cannot be represented internally . In this respect, it is a physical Navigation as Computation 17 ,. ~:,;,I~ f9n 2.18 Anastrolabe , whichsuperimposes several kinds of sml Cblreto CI8Ite a celestialcomputer . residuum of generations of astronomical practice . It is asedimentation of representations of cosmic regularities . The astrolabe also enables its user to predict the positions and movements of the sun and the stars: Because the asmolabe can be set to show the positions of these heavenly bodies at different times of day or night , on different dates and or different latitudes , the instrument is also a computer , serving to solve problems concerning the positions of the Sun and stars at any given time . (National Maritime Museum 1976) Any map of the heavens can capture the relationships among the stars. The astrolabe goes further . The physical structure of the Chapter2 . moving parts of the instrument captures regularities in the movements of the heavens and the effects of latitude and time on the observations of the heavens. Thus , the astrolabe is not just amemory for the structure of the sky; it is also an analog computer . The major components of an astrolabe are the mater , the limb , the plate , and the rete. The mater is the framework that holds the other pieces together . The limb is a circular scale around the perimeter of the mater . The limb is inscribed with a 3600 scale and/or a 24-hour scale. In either case, the limb is a representation of the structure of sidereal time . Each astrolabe is really a kit that can be assembled differently according to the circumstances of its use: As the configuration of the celestial coordinates changes according to the latitude of the obse I Vera set of removable plates - sometimes as many as six, engraved on both sides- is usually supplied , fitting into the hollow of the mater , so that the user can select the plate most appropriate to his own latitude . (National Maritime Museum 1976: 14) The interchangeable plates capture regularities in the effects of observer latitude on the relations of the celestial coordinates to the local horizon . Of course, it is not possible to provide a plate for every observer latitude , since latitudes are infinite in number . The plates provide a coarse discrete representation of the effects of latitude . Even with a large number of plates , the representation of observer latitude will be approximate most of the time . The rete captures the locational relationships of the stars to one another and that of the sun to the stars. The assembled astrolabe brings these three kinds of structure (and much more) into coordination just the right way so that the interactions of these variables can be control led in the manipulation of the physical parts of the instrument . An astrolabe can be made of durable materials because the regularities it captures change only very slowly . The variables that do change, observer latitude and time of observation , are represented in the physical structure of the astrolabe either by changeable parts (plates for each latitude ) or by changeable relations among parts (the rotation of the rete about the axis with respect to the plate and the limb ). The constraints of the represented world are thus built into the physical structure of the device . The astrolabe is a manipulable model of the heavens- a simulator of the effects of time and latitude on the relationships of the heavens to the horizon . The astrolabe is an early Navigation as Computation . example of a general trend toward the representation and solution of computational problems via physical manipulations of carefully constructed artifacts . THE COMPASS ROSEAND RECKONING THE TIDES Frake (1985) provides an especially interesting example of the ways in which a variety of kinds of structure are combined in a single artifact to create a computational system. Frake is interested in what Northern European sailors knew about the tides and how they went about knowing what they know . Although he is interested in the tides , his account begins with the so-called wind rose: The schema of directions . . . resulted from a successive division of the quadmnts of a horizon circle formed by n(:lrth -south and east- west lines into 8, 16, and finally 32 named (not numbered ) points . . . . Similar schemata for segmenting the circle of the horizon with invariant directional axes chamcterize all known early seafaring traditions : those of the Pacific , the China Sea, the Indian Ocean and Europe . In the various traditions , compass directions could be thought of as, and named for , star paths (as in the Pacific and the Indian Ocean) or wind directions (as in island southeast Asia and Europe). In all cases, the compass rose provided an invariant representation of directions which were, in fact , determined at sea by a variety of means : the sun, stars, winds , swells, landmarks , seamarks, sea life and , in later times, the magnetic needle . (Frake 1985: 262) The wind rose is an ancient schema that , for most of its history and in most places, had nothing in particular to do with representing knowledge of the tides . In the Mediterranean , for example , the tides vary so little that mariners can safely disregard them . In Northern Europe , by contrast , tidal variations are large, and the ability to predict the tides is of great value to mariners . The use of the medieval compass rose in the prediction of tides is a fine example of the empirical construction of an artifact in the absence of a theory of the phenomenon it permits navigators to predict . The compass rose as a schema for the expression of directions was appropriated as a schema for the representation of time as well (see figure 2.19): In whatever manner time was determined at some moment , it was thought of and expressed as a compass bearing . The sun bears Chapter2 1. Midnight 11 11 : 0 : 458 10:3Op 15p bYE bYE N N bywN bywN N N 1 : 308 9:45p NNW 2:158 Solartimeof NW by N by N 9:00p NE : 008 hightide \. 3 ' ---...... 8 :15p E NWbY 3:458 I E : 30a 7: 4 ~ E FIVedays WNW past I 5:158 full moon 6:45pWby N IEby N t 6:008 6:00p. W E t 'Eby S 6 :458 5:15pWby S , '\ ESE7:3Oa - - 4:3Opwsw SWby r 3:45p Lunartime of SE hightide 3:00p 9:00a (establishmentof : the 2 port) 15p SSE 1 : by W W 3Op 3Op S S SbyE 10 : 30a S S : 12 11:158 45p Noon ~ 2.18 Computingthe tide fromthe su~ mp O Iition of ternporlilandnwkson the compassrose . south at noon . It was therefore thought of as bearing north at midnight , east at 6 aim . and west at 6 p .m . Only the first of these bearings is of practical daily use in northern Europe for determining time . The other bearings were ways of expressing time . (Frake 1985: 264) Here we have the superposition of two kinds of structure : the temporal structure of the 24-hour solar day on the 32-point compass rose. This yields a set of correspondences between direction and solar time . If the bearing of the sun is an expression of solar time , the bearing of the moon can likewise be seen as an expression of lunar time . The tides result from the gravitational pulls of the moon and the sun. The effects of the moon predominate . Although the tide does not simply follow the moon in any obvious manner , the phase of the tide at any particular place is always the same when the moon Navigation as Computation 101 is at any given bearing . That is, for any particular location , the high tide always comes at a particular lunar time . Medieval mariners noticed this fact: Medieval sailing directions , and presumably the memories of sai- lors before written directions , specify the tidal regime of a given place by stating the lunar time , named as a compass bearing , of a given state of the tide , usually high . (Frake 1985: 265) With both solar and lunar time superposed on the compass rose, the relationships between solar time and lunar time can be expressed as directional relationships . A sailor who knows the lunar time of high tide for a given location can use the superposed lunar and solar time representations to compute the solar time of high tide . For example , if it is known that at a given location the high tide will occur when the moon bears WSW , " the sailor has to determine the solar time corresponding to WSW " moon on a given date and also calculate the state of the tide at any other solar time . It is in the solution of this problem that the compass rose as cognitive schema shows its merits . (Frake 1985: 265) The simplest case occurs when the phase of the moon is new . In that case, solar time and lunar time are the same, and the time of high tide will be when the moon (and therefore the sun as well ) bears WSW . That is, high tide will come at 4:30 p.m. If the phase of the moon is other than full or new , the sailor will have to first determine the relation of solar time to lunar time in order to compute the time of the high tide . It just so happens that dividing the 24- hour day into 32 equal intervals yields intervals of 45 minutes each: Each day, lunar time , and the tide following it , lags behind the sun by about 48 minutes . Our compass points divide time into 45-minute intervals , close enough to 48 for tidal calculation . (Frake 1985: 265) Suppose a sailor finds himself approaching this harbor five days past the full moon . Since the moon and the tide lag 48 minutes behind " the sun each day, we can count five points of the compass " past WSW to NW by W , a point which marks the solar time of 8:15 (Frake 1985: 265). In this way , the sailor can compute the solar time of the high tide (and therefore the other tides as well ) by knowing Chapter2 1-. the phase of the moon and the establishment of the port (figure 2.19). The tidiness of the compass rose as a representation of these relationships is an entirely fortuitous property of the mapping of the 24-hour day onto the 32 points of the compass rose. The segmentation of the compass rose into 32 points and the segmentation of the day into 24 hours arose independently . Their relationship just happens to map approximately onto the 48-minute daily lag of the moon behind the sun that results from the relation of the 29.5-day lunar cycle to the 24-hour day. The superposition of the scheme for the 24-hour day on the scheme for the 32-point wind rose yields a system of temporal and spatial landmarks on which the correspondences of the states of the tide and time can be imagined and represented . This is reminiscent of the superposition of temporal and directional landmarks that the Micronesian navigators use to compute their progress on a voyage. Frake noted the abstract similarities of the two systems: It is the relationship between determining direction and determining time that makes the use of a single schema, the compass rose, appropriate for representing both direction and time . But the compass rose is not a time -finding instrume.nt . It is a very abstract model , a cognitive schema, of the relations of direction to time , of solar time to lunar time , and of time to tide . It is an etak of medieval navigation . (Frake 1985: 266) ' Frake s comparison of the compass rose used to compute tides to the Micronesian concept of etak is based in the abstract properties of both as organizing schemata. I believe that the links are even stronger in that both systems achieve their computational power by superimposing several kinds of representational structure on a single framework . Both of these devices- the astrolabe and the compass rose as tide comput ~r- involve the creation of physical artifacts whose structures capture regularities in the world of phenomena in such a way that computations can be performed by manipulating the physical devices. It should be noted , however , that the use of the compass rose as a tide computer is a bit more like the Micronesian navigation case, in that an important part of the structure is not explicitly represented in the artifact itself but is instead supplied by the situated looking of the navigator . Navigation as Computation 103 M88irlment81d ' - - .0l OI111 of DIg I IICor ~ tation A second clear difference between the Micronesian and Western navigation traditions is the reliance of the latter on measurement and digital computation . This difference is apparent in the history of the chip log. THECHIP LOG The spread of the use of the chip log in about 1600 marks an important turning point in the history of Western navigation . Before this , European navigation was based primarily on analog computations . The log gave rise to a computational process that begins with - - analog to digital conversion , which is followed by digital computation - - , then either digital to analog conversion for interpretation or - - digital to analog conversion followed by analog computation . Western navigators have been practicing this style of navigation for less than 400 years. The chip log is a simple analog-to-digital converter that converts the rate of travel of a ship through water into a number by making a direct measurement of the distance the ship moves in a given unit of time . A panel of wood , called the chip , is tied to a line and thrown over the side of the ship (figure 2.20). It remains stationary F9ire2 .20 A ch~ log. (FromMaloney 1985 .) Chapter2 104 in the water while the ship sails away from it . The line attached to the chip is allowed to payout , and the amount of line that pays out in a given period of time is the distance the ship has traveled in that same period . Since speed is distance per unit time , this distance is, by definition , directly proportional to the speed of the ship . In the early days of the use of the chip log, the interval of time was measured by the duration of a spoken prayer . Later , to increase accuracy , a sand glass was used instead . To measure the distance covered , one could reel the line back in and then measure the amount that had paid out during the given interval . It would be exceedingly difficult for a single person to perform this procedure with accuracy . I have witnessed the use of a traditional chip log aboard the restored late-nineteenth -century Swedish cargo schooner Westkust . The procedure requires three people working in close coordination . One manages the chip and the line , throwing the chip overboard , letting the line run through his fingers " " , and calling out when the end of the stray line has passed his hand . A second person inverts a sand glass when the first indicates that the measured portion of the line is now streaming out , and calls out when the sand has run out . When the time is up , the first person grips the line and stops paying out the line . This stop- ping is assisted by a third person , who has been, up to this point , holding the spool on which the line is wound so that it can flow out smoothly . The line is dressed with tassels hanging from the knots so that the number of each knot can be discerned at a glance. The number of the knot nearest to hand is noted , and the line is pulled in and wound onto the spool . Columbus did not mention the use of a chip log, although his logbooks do contain entries recording speeds. It is assumed that he either estimated his speed by eye or used a precursor of the chip " log that involved dropping a piece of wood into the water and " timing the passage from bow to stem (Fuson 1987: 44). As with the early chip logs, the interval of time was measured by the recitation of a chant or a prayer . The first certain use of the chip log ' was on Magellan s voyage in 1521. The use of the log or any other technique based on the distance covered by the ship requires both a consistent unit of distance and a means of reliably measuring distance in that unit . This was accomplished by preparing the log line in a special way : Navigation as Computation 1~ By [1633] it had become the general practice to mark the log line so as to facilitate the calculation of speed. This was done in the following way. If a half minute glass was used then the length of line necessary to indicate a speed of one mile (of 5000 feet) per hour was (30 x 5000ft )/ (60x 60) or 41ffeet . ln other words, at one mile per hour the ship would advance , and the line would run out 41f feet in 30 seconds. The line was then divided as follows : From 10 to 20 fathoms , depending on the size of the ship , were allowed as " " stray line next to the log chip , to ensure it being clear of the effect of the wake. The end of this stray line was marked either by a knot or a piece of red or white rag, and then from there the line was divided into sections of 41f feet or 42 feet , each section being marked by a knot in the line . Thus came into being the term known ' as the measure of a ship s speed in nautical miles per hour . (Hewson 1983: 160) Even with these refinements , the chip log was not a very accurate instrument . Many things could induce errors in the readings . The friction of the spool , shrinkage of the rope , the surge of the ship working in steep seas, the effects of currents , and the yawing of the ship with a swell on the quarter were among the many things that could cause significant errors . For a navigator relying on a log, there is no choice but to expect error and attempt to allow for it . Just as a carpenter would rather err on the long side in cutting a piece of wood (so that any error can be corrected with minimal waste of material ) a navigator prefers to overestimate the distance sailed in order to avoid an unexpected landfall . If an error is made, it is better to have overestimated the distance sailed , so that the problem can be corrected without losing the ship . Log lines can shrink with use, so it is important to check the " length of the segments between the knots . This was facilitated in most ships by having permanent marks of nails driven into the " ' deck (Hewson 1983: 166). Decks don t stretch and shrink as ropes do. Putting the calibrating nails into the deck is a way of creating a memory for the lengths between knots in the log line in a medium that has physical properties that match the computational needs of the task. In this case, the marks on the deck are a memory for distance . In the late eighteenth century many attempts were made to develop more accurate ways to measure speed or distance run Chapter2 1. through the water . These included the taff-rail and paddle -wheel logs (Hewson 1983). Although the details of their implementation varied , these were all simple analog-to-digital converters that stood in the same relation computationally to other navigation tools that the chip log had. The importance of the chip log is that it changed the way navigation was done. Rather than knowing a journey should take some number of days and counting days until the required number had elapsed, a navigator using a chip log used the concept of distance between points and the integration of speed over time to determine the distance covered by the ship . Having created a digital representation of speed, the chip log created a need for a method of calculation that could operate on that representation to tell the navigator what he needed to know . The chip log and its descendants are among the many measuring instruments that entered the navigation tool kit during the European expansion . Others include a succession of instruments for measuring the altitudes of stars (astrolabe, quadrant , cross-staff, sextant), range-measuring instruments , instruments to measure bearings , azimuths , and courses, and instruments to measure depths . All of these are analog-to-digital converters . All of them create representations that are subsequently processed using a special arithmetic technology in order to produce information that is of use to the navigator . Consider the enormous importance of common logarithms . With a table of logarithms , one can transform multiplication and division into addition and subtraction . That is, when numerical values are expressed as logarithms , the complex typo graphic operations required for multiplication and division (the algorithms of place- value arithmetic ) can be replaced by a simpler set of typo graphic operations that implement addition and subtraction . Speaking of - Edmund Gunter (1581 1626), Cotter (1983a: 242) says: He introduced the first tables of logarithmic trigonometrical functions , without which a seaman would find almost insurmountable ' difficulty in solving astronomical problems . It was Gunter s Tables, published in 1620, that paved the way to the new phase of " " arithmetic navigation . Armed with the new logarithmic canon, a navigator who memorized the necessary rules could solve nautical astronomical problems with relative ease. Navigationas Computation 107 But the seamen of the time found even the simplified calculations daunting , so Gunter designed a ruler with a number of scales: Among these are a logarithmic scale of natural numbers , logarithmic " scales of sines, tangents and versines; . . . and a meridian " ' line to facilitate the construction of sea charts on Wright s projection " " . . . . With the advent of arithmetical navigation , in which Gunter played the dominant role , the common log for measuring a ' ship s speed became commonplace . To the careful seaman using a Gunter scale the proportional problem of finding speed was mechanical and, therefore , trivial . (ibid .) The predecessor of the slide rule is apparent here. In fact, it appears that two of Gunter 's scales were sometimes " laid down on rulers to " slide by each other (Oxford English Dictionary , 1971). Again we have an artifact on which computations are performed by physical manipulation . However , there is an important difference between ' the astrolabe and Gunter s scale in this regard. In both cases the constraints of a represented world are built into the physical structure ' of the device , but in the case of Gunter s scale the represented world is not literally the world of experience . Instead it is a symbolic world : the world of logarithmic representations of numbers . The regularities of relations among entities in this world are built into the structure of the artifact , but this time the regularities are the syntax of the symbolic world of numbers rather than the physics of a literal world of earth and stars. The representation of symbolic worlds in physical artifacts , and especially the representation of the syntax of such a world in the physical constraints of the artifact itself , is an enormously powerful principle . The chip log and ' Gunter s scale are representative elements of a cognitive ecology based on measurement and digital computation . TheCh8t . 8 Modelof lie World The navigation chart - perhaps the best available example of the crystallization of practice in a physical artifact - is intimately involved in the prototypical cycle of measurement , computation , and interpretation that characterizes so much of Western navigation . These characteristics of the chart will be developed in much more detail in the coming chapters . At this point it is useful to examine another contribution of the chart that marks one of the most im - Chapter2 1~ portant elements of the Western conception of navigation . The chart , by virtue of its interpretation as a model of an expanse of actual space, encourages a conception of a voyage as sequence of locations on the chart . Descriptive sailing directions were the principal navigation aids up until the end of the eighteenth century . These documents describe to the sailor how to proceed with the voyage and what he can expect to see. Then , with the continued improvement of survey techniques and the increasing range of areas accurately surveyed , sailing directions were supplanted by the pictorial chart . This marks an important change in perspective . Where the sailing directions presented the world from the perspective of the deck of the - ship , the coastal chart presented the world from above from a " ' " virtual perspective (a bird s-eye view ) that navigators would never actually experience . Modem navigators may take to the air ' and adopt something very like a bird s-eye view , but this is not in fact the perspective presented by the navigation chart . The navigation chart presents the world in a perspective that can never be achieved from any actual viewing point . A chart must be more than an accumulation of observations . The structure of the chart is crucial (Cotter 1983b). The importance of the compass in the actual practice of navigation was paralleled by its contribution to the quality of chart production . The compass made it practical to make accurate charts . It was possible before (by means of the stars) to get directions for bearings and courses, but not nearly so conveniently . Even when a compass was used, serious problems in chart construction remained . For example , early charts of the Mediterranean showed a pronounced upward tilt in the eastern end. This tilt was produced by the difference in magnetic variation between the western and eastern reaches of the Mediterranean Sea. If the cartographer uses a magnetic compass to make the chart , and the navigator uses a magnetic compass to determine courses, and if both compass es show the same errors in the same places, why would anyone care and how could anyone ever notice that the charts put the land in the wrong places? TAKING THE MEASURE OF THE EARTH The distortions in charts produced by changes in magnetic variation became an issue when the chart became a point of articulation between the measure of the earth and celestial observations . In or- Navigation as Computation 1. der for the effects of distance covered on the face of the globe to be reconciled with the attendant , and also measurable, changes in latitude (a relation to the celestial sphere), the unit of measure for distance had to be grounded in the measure of the earth itself . That ' is, a degree of arc on the earth s surface is a particular distance , and navigators wanted to be able to combine and interrelate measurements made in terms of distance traveled with measurements of latitude . For example , if I am currently 20 south of my home port , sailing north , how far, in units of distance on the surface of the ocean, must I sail to arrive at the latitude of my home? The question is: How much linear distance on the surface of the earth corresponds to a degree of arc on the same surface? As we saw in the discussion of the historical changes in the length of the nautical mile , establishing a standard that permit ted the chart to be a point of articulation between the measure of the earth and the measure of the heavens was no simple task: - The north up convention is clearly related to the concept of defining position in terms of latitude and longitude . For coastal navigation this concept is of no consequence: a coastal navigator is ' interested in defining his ship s position not in terms of these spherical coordinates , but in terms of bearings and distances from prominent landmarks of hazards such as rocks and shoals. Early coastal charts , therefore, (and with good reason) were orientated relative to the run of the coast rather than to the compass. (Cotter 1983b: 256) The modem chart incorporates the global convention of north -up depiction of a plane surface having a discrete address in terms of latitude and longitude for every location . This global framework permits the combination of any number of observations from any number of locations . With this scheme it is possible to compute the relationship between any two locations on earth even though that relationship has never been measured. The virtual perspective created in the chart does not privilege any actual perspective . A navigation chart is a representation that is equally useful (or not useful ) from any actual perspective . It attempts neutrality with respect to the perspectives from which the world will be seen by navigators . Since in the Western tradition nearly all navigable space is represented from this virtual perspective , it is from this virtual perspective that voyages come to Chapter2 110 be conceived . We imagine the voyage as the movement of our ship over a stretch of water . There is the ship , and there we are, like tiny imagined specks on the tiny imagined ship that is moving in our ' mind s eye across the expanse of paper that represents the water between origin and destination . Yet there are moments in which this perspective does not serve the needs of the navigator , as when one attempts to determine what the land depicted should look like from the perspectives that are actually achieved in ships . Here coastal profiles may be included . There is a problem at the moment in which one moves conceptually " " " " from being on the chart to being in the world . The coastal profile is a concession to this problem . The first coastal ' profiles appeared in 1541, in Pierre Garcie s book Le grand routier . Coastal profiles are representations that privilege particular perspectives that the chart makers anticipate will be encountered often by users of the chart . The common framework of locations also permits the superposition of a wide variety of structures . In addition to the obvious boundaries of bodies of water and land , the locations of cultural features and of geographical features (both above and below water ) are depicted . This superposition of these structures , which underlies much of the computational power of the chart , is so obvious as to go unnoticed by virtually all the users of the chart . Soundings were first shown reduced to a standard half -tide datum in 1584, in ' Janzsoon Waghenaer s Speighel der Zeevaert. SOCIAL,PROBLEMSOFCHART CONSTRUCTION The birth of astronomical navigation was much less a scientific problem than a question of organization . Jean II of Portugal had the great merit to have known - before any other head of state- to organize the technical exploitation of the theoretical knowledge of his epoch .- BeaujouanScienceLivresque et Art Nautique au xve Slecle; cited in Waters 1976: 28 (translation byE .H.) There is a great deal of knowledge embodied in any navigation chart . To add a new feature to a chart , one must determine its relationship to at least two other features. Since a chart implicitly represents a spatial relationship between the members of every pair of features depicted , any the new feature acquires relationships to all the other features on the chart - not just the ones that were Navigation as Computation 111 used to establish its location . If the number of relationships depicted is a measure of the knowledge in a chart , there may be more knowledge in a chart than was put into the chart . In fact , most of the relationships depicted on any chart have never actually been measured. Even so, a great many observations are required in order to construct a useable navigation chart . A navigation chart represents the accumulation of more observations than anyone person could make in a lifetime . It is an artifact that embodies generations of experience and measurement . No navigator has ever had , nor will one ever have, all the knowledge that is in the chart . The really difficult technical problem in the production of charts is the collection of reliable information . (See Latour 1987 on centers of calculation .) Compare the problem faced by the Portuguese during their expansion with that faced by the Micronesians . Every Micronesian navigator knows the courses and distances between all the islands in his sailing range- including , as we have seen, courses between islands that have not been visited for many generations . How could a Micronesian navigator come to have this knowledge ? Clearly , it is acquired over generations , and what any navigator knows is much more than could be learned by direct observation . The knowledge is a compilation of the experiences of many navigators - some of whom , one must assume, set out on voyages of discovery , knowing which way they were sailing , and how to get home , but not what they would find . Over the years the knowledge accumulated , expressed in the framework of star courses and etak images. Today the knowledge of a Micronesian navigator exceeds what could be acquired by direct observation , but it does not exceed what could be remembered by one individual . The world of the Portuguese fleet in the early fifteenth century was much larger than one group of islands . The total knowledge of the world not only exceeded what could be observed by any individual , it exceeded what could be known by any individual . Like the Micronesians , the Portuguese needed a consistent set of techniques for making observations and a representational framework in which all observations could be expressed. They also needed to train a large number of observers in these techniques so that the experiences of all of them could accumulate in acom - mon store. This was the creation of an enormous system for gathering and processing information - a cognitive system of many Chapter2 112 parts that operated over many years to create a collection of representations of the spatial organization of the surface of our planet (Law 1987). TheComputational Ecology of NavigationTools The mutual dependencies among the various instruments and techniques is clearly visible in the history of navigation . Even though the chip log was available for use in the sixteenth century , for example , it was not generally adopted until the middle of the ' seventeenth . Why weren t sailors using the log more widely ? Because they had no convenient way to carry out the computations required to turn the readings gained from the log into useful information ' about the ship s position . Why was there , before 1767, no nautical almanac giving the positions of the stellar sphere, the sun, the moon and the planets ? Astronomy was certainly advanced enough to provide these data. The answer is that these data are useless for marine navigation in the absence of an accurate way to determine time at sea. The need was well known , and in 1714 the English Parliament passed an act " providing a Pub lick Reward for such person or persons as shall " discover the Longitude at Sea. The reward went unclaimed until 1762, when Harrison constructed a chronometer that would work reliably at sea (Taylor 1971: 261). The nautical almanac soon followed . Before seagoing chronometers were perfected , there was little incentive to develop better sextants. At the equator , an error of one minute in time produces an error of 15 miles in east-west position . Since the earth turns on its axis one degree of arc in 4 minutes of time , there is no utility in having an instrument that can measure celestial angles-even to the nearest degree unless it is coupled with a chronometer that is accurate to within 2 minutes . Thus , both the development of the sextant and the development of accurate navigation tables were arrested by the lack of the chronometer . Both were technological possibilities before the development of the chronometer , but there was no use for them until time could be reckoned accurately . Similar dependencies can also be seen in the history of the chart and the plotting tools . Charts were in wide use by the thirteenth century , but the most basic of plotting tools - the parallel rule - was not invented until the late sixteenth century ( Waters 1976). Navigation as Computation 113 Why ? Because a straight line has no special meaning on an early chart . Not until the Mercator projection did a straight line have a computationally useful meaning . But the earliest Mercator chart came with no explanation . It is unlikely to have been used at sea ( Waters 1976). Navigators needed instruction in the use of exotic technologies . (Note : For some time , the Mercator projection was known to the English -speaking world by the name of the man who published an English version of it in 1599: Edward Wright .) The early astrolabes and quadrants were university equipment . ' Ordinary seamen couldn t use them . The tools had to be simplified , and there had to be instructions in their use: By themselves these instruments (quadrant and astrolabe) were, of course, powerless . The mere fact of sighting a heavenly body through pinholes of an alidade had nothing per se to do with navigation . That sighting , or the reading that corresponded to it , had to undergo a number of complex transformations before it could be converted into a latitude . The construction of a network of artifacts and skills for converting the stars from irrelevant points of light in the night sky into formidable allies in the struggle to master the Atlantic is a good example of heterogeneous engineering . (Law 1987: 124) Sometimes, as the nature of the practice has changed, the role of particular instruments has changed. For example , the astrolabe was originally used both to measure the altitudes of celestial bodies and to predict the altitude and azimuth of a star. The observation - making duties were subsequently taken over by the quadrant , then by the cross staff, and finally by the sextant. The function of computing the expected altitudes and azimuths of stars was taken over by a complex set of tables. Even though the quadrant and the cross staff were eventually replaced by the sextant (which is much easier to use), their ancestor, the astrolabe, survives as the modem star finder . It is now usually made of plastic instead of brass, but it is easily recognizable . A star finder is not considered accurate enough for the purposes of computing expected altitudes , but it is used to set the sextant before making the observation . It is used to get the setting of the precision instrument into the right neighborhood . It has been moved to a new job in the navigation process. In attempting to understand the history of navigation from a cognitive perspective , it is important to consider the whole suite of instruments that are used together in doing the task. The tools of Chapter2 114 navigation share with one another a rich network of mutual com- putational and representational dependencies . Each plays a role in the computational environments of the others , providing the raw materials of computation or consuming the products of it . In the ecology of tools , based on the flow of computational products , each tool creates the environment for others. This is to see in the ' easy history of the physical tools , but the same is certainly true of the ' mental tools that navigators bring to their tasks. Frake s compass rose is there for all to see, but it becomes a tide computer only in interaction with the establishment of the port and with a particular way of seeing the circle of directions as a representation of the temporal relationships of the periodic cycles of the sun and the moon . Every argument showing why a particular tool is easy to use is also an argument showing why both internal and external tools are part of the very same cognitive ecology . It is a truism that we cannot know what the task is until we know what the tools are. Not only is this true of both internal and external tools , it is also true of the relationships among them . TheTr8l Spire I ICyof C ~ Rep.~ ~ ~ HowW . Fallto See~ (Oln - Theirs) I have presented this comparative and historical treatment to remind us all that the ways we have of doing things , the ways that seem to us to be natural and inevitable or simply the consequences of the interaction of human nature with the demands of a given task, are in fact historically contingent . As Benedict (1946: 14) notes, The lenses through which any nation looks at life are not the ones another nation uses. It is hard to be conscious of the eyes through which one looks . Any countzy takes them for granted , and the mcks of focussing and of perspective which give to any people its national view of life seem to that people the god-given arrangement of the landscape . In any matter of spectacles, we do not expect the man who wears them to know the formula for the lenses, and neither can be expect nations to analyze their own outlook upon the world . Navigationas Computation 115 Of all the many possible ways of representing position and implementing navigation computations in the Western tradition , the chart is the one in which the meaning of the expression of position and the meaning of the operations that produce that expression are most easily understood . As was noted above, lines of position could be represented as linear equations , and the algorithm applied to find their intersection could be that of simultaneous linear equations . As a physical analog of space, the chart provides an interface ' to a computational system in which the user s understanding of the form of the symbolic expressions (lines of position ) ' is structurally similar to the user s understanding of the meanings of the expressions (relations among locations in the world ) (Hutchins , Holian , and Norman 1986). In fact, the similarity is so close that many users find the form and the meaning indistinguishable . Navigators not only think they are doing the computations , they also invest the interpretations of events in the domain of the representations with a reality that sometimes seems to eclipse the reality outside the skin of the ship . One navigator jok - ingly described his faith in the charted position by creating the " following mock conversation over the chart : This little dot right here where these lines cross is where we are! I don 't care if the bosun says we just went aground , we are here and there is plenty of " water under the ship here. For the navigator , the ship is where the lines of position intersect . It is really astonishing how much is taken for granted in our current practice . The difficulties that were overcome in the creation of all these techniques , and the power they provide relative to their predecessors, are not at all apparent to the modem practitioner . Only when we look at the history can we see just how many problems had to be solved and how many could have been solved differently in the course of the development of the modem practices . A way of thinking comes with these techniques and tools . The advances that were made in navigation were always parts of a surrounding culture . They appeared in other fields as well , so they came to permeate our culture . This is what makes it so difficult to see the nature of our way of doing things and to see how it is that others do what they do. We see in the divergence of these traditions not just the development of the tools of measurement , but a passion for measuring and a penchant for taking the representation more seriously than the thing represented . Chapter2 118 While all navigation computations seem to be describable by a small number of abstract principles , there is great variation in the representational systems and concomitant algorithmic procedures that may be employed to organize the computations . The actual devices and process es in which these representations and algorithms are implemented have a complex evolutionary history . In the next chapter we will consider in much greater detail the implementation of the computations of Western navigation . 3 71 Ie'm' 8I ' Bltaf D' of C G'te I', . . " PIGfa. . , .